the 2012 labour market reform in spain: a preliminary assessment
TRANSCRIPT
![Page 1: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/1.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT December 2013
![Page 2: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/2.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN:
A PRELIMINARY ASSESSMENT
DECEMBER 2013
![Page 3: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/3.jpg)
![Page 4: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/4.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 3
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
FOREWORD
This report provides an initial evaluation of the comprehensive reform of the Spanish labour market
undertaken in 2012. The report was commissioned to the OECD by the Spanish government and it
complements the evaluation of the 2012 labour market reform undertaken by the Labour Ministry
(Ministerio de Empleo y Seguridad Social (MEySS)) and presented in September 2013.
The evaluation presented in this report should be considered preliminary and mainly confined to the
short-time impact of the reform, given the fact that only a short amount of time has passed since the reform
was undertaken and the complexity of assessing the impact of such a comprehensive reform.
The objective of this report is to describe the key components of the 2012 reform and place them in
the context of the evolution of labour market institutions in other OECD member countries, with a
particular focus on collective bargaining and employment protection legislation. The report also assesses
the impact of the reform on the ability of firms to adjust wages and working time to cope with demand
shocks (so-called internal flexibility), as well as the flows in the labour market for different types of
contracts and the overall duality of the Spanish labour market. The report also considers what
complementary reforms would be required to improve the effectiveness of the labour market reform, in
particular in the area of active labour market policies.
The work on this report was carried out by the Directorate for Employment, Labour and Social Affairs
(ELS) of the OECD. The report was prepared by Andrea Bassanini and Josep Mestres, with statistical
assistance from Thomas Manfredi. It benefited from the many useful comments by Stefano Scarpetta
(ELS Director), Mark Keese (Head of the Employment Analysis and Policy Division) as well as by staff in
the OECD Economics Department.
The co-operation of Spanish Authorities in providing the data used in this report is gratefully
acknowledged. José Ignacio García Perez kindly performed the estimations of competing-risk hazard
models on social security data. Moreover, staff in the Spanish Labour Ministry provided useful comments
on previous drafts. Nonetheless, the OECD Secretariat bears full and sole responsibility for this report.
This work is published on the responsibility of the Secretary-General of the OECD. The opinions
expressed and arguments employed herein do not necessarily reflect the official views of the Organisation
or of the governments of its member countries.
![Page 5: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/5.jpg)
![Page 6: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/6.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 5
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
TABLE OF CONTENTS
EXECUTIVE SUMMARY ............................................................................................................................. 7
INTRODUCTION ......................................................................................................................................... 11
1. Key components of the 2012 reform ..................................................................................................... 12 2. The labour market reform in international perspective .......................................................................... 14
2.1 Collective bargaining ....................................................................................................................... 14 2.2 Employment Protection Legislation ................................................................................................. 19
3. The labour market impact of the 2012 reform ....................................................................................... 25 3.1. Labour costs .................................................................................................................................... 26 3.2. Changes in worker flows ................................................................................................................. 29 3.3 Simulating the long-run impact of the 2012 reform on productivity and economic growth ............ 40
4. Interactions with other policies .............................................................................................................. 41 5. Conclusions and recommendations ........................................................................................................ 44
References ....................................................................................................................................................... 47
ANNEX: EMPIRICAL FRAMEWORK AND DETAILED ESTIMATION RESULTS ............................. 53
Tables
Table 1. Structure of collective bargaining systems in Europe before the crisis ....................................... 15
Table A1. The effect of the reform on year growth of labour costs........................................................... 56 Table A2. The effect of the reform on the probability of having tenure shorter than three months .......... 58 Table A3. The effect of different placebos on the probability of having tenure shorter
than three months ....................................................................................................................................... 58 Table A4. The effect of the reform on new permanent contracts .............................................................. 60 Table A5. The effect of different placebos on new permanent contracts .................................................. 62 Table A6. The effect of the reform on new permanent contracts, by type of transition ............................ 62 Table A7. The effect of the reform on inflows onto unemployment benefits due to dismissal ................. 64 Table A8. The effect of the reform on separation rates at the establishment level .................................... 67 Table A9. The effect of the reform on Transitions out of Unemployment into temporary
employment and into permanent employment ........................................................................................... 70 Table A10. The effect of the reform on Transitions out of Unemployment into temporary
employment and into permanent employment, by personal characteristics .............................................. 72
Figures
Figure 1. Percentage of firms subject to a mechanism of adjusting base wages to inflation
before the crisis .......................................................................................................................................... 16 Figure 2. Evolution of unit labour costs in Euro countries, 2000Q1-2008Q1 ........................................... 16 Figure 3. Adjustment strategies to adverse shocks used by European firms before the crisis ................... 18 Figure 4. Protection of permanent workers against individual and collective dismissal,
2008 and 2013 ............................................................................................................................................ 20
![Page 7: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/7.jpg)
6 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Figure 5. Employment protection for permanent workers: Spain vs. OECD ............................................ 20 Figure 6. Incidence of temporary employment, 2007 and 2012 ................................................................ 24 Figure 7. Evolution of unit labour costs in selected European countries, 2008-2013................................ 27 Figure 8. Evolution of the wage and salary component of unit labour costs in Spain, by industry,
2008-2013 .................................................................................................................................................. 27 Figure 9. Estimated impact of the 2012 reform on the year-on-year growth of business-sector
labour costs ................................................................................................................................................ 28 Figure 10. The evolution of hiring rates in the non-agricultural business-sector, 2008-2013 ................... 30 Figure 11: Average proportion of unemployed individuals leaving unemployment to permanent
employment ............................................................................................................................................... 31 Figure 12: Exit from unemployment by type of contract and unemployment duration ............................. 32 Figure 13: The effect of the reform on quarterly separation rates at the establishment level .................... 33 Figure 14: The effect of the reform on separation rates at the establishment level by type ...................... 34 Figure 15. Share of inflows onto unemployment benefits due to collective dismissal in total
inflows due to dismissal, 2007-2013 ......................................................................................................... 35 Figure 16. Monthly inflows into permanent contracts, including contract conversions, 2006-2013 ......... 37 Figure 17. Exit from unemployment to permanent employment by firm size ........................................... 38 Figure 18. Average of net replacement rates over the first 60 months of unemployment, 2011 ............... 43 Figure 19: The Beveridge curve in Spain .................................................................................................. 43 Figure A1. New hires with a permanent contract in firms with 50 or fewer employees, 2006-2013 ........ 61 Figure A2. New hires with a permanent contract in firms with more than 50 employees, 2006-2013 ..... 61 Figure A3. Share of inflows onto unemployment benefits due to collective dismissal in total
inflows due to dismissal, 2007-2013 ......................................................................................................... 64 Figure A4. The effect of different placebos on the share of inflows onto unemployment benefits
due to collective dismissals ........................................................................................................................ 65 Figure A5: Exit from unemployment by type of contract and unemployment duration ............................ 70 Figure A6: Exit from unemployment conditional on duration smaller or equal than 6 months,
by type of contract and establishment size ................................................................................................ 71
Boxes
Box 1. Estimating the impact of the 2012 reform using regression-discontinuity models ........................ 25
![Page 8: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/8.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 7
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
EXECUTIVE SUMMARY
This report provides an initial evaluation of the impact of the comprehensive reform of the Spanish
labour market undertaken in 2012. This assessment is complementary to the evaluation performed by the
Ministerio de Empleo y Seguridad Social (2013). Nonetheless, the findings presented here should be
considered as preliminary, given the short amount of time that has elapsed since the reform was introduced
and the complexity of assessing the impact of such an extensive reform.
The reform has promoted the internal flexibility of firms…
One of the key elements of the reform is the greater priority given to collective bargaining agreements
at the firm level over those at the sectoral or regional level and the greater possibility given to firms to
opt-out from a collective agreement and adopt internal flexibility measures to limit job destruction. In
addition, the extension of collective bargaining agreements after their end date if no new agreement is
reached is now limited to a maximum period of one year. Dismissal regulations have also been modified,
redefining the conditions for fair dismissal, reducing monetary compensations in the case of unfair
dismissal and eliminating the requirement of administrative authorisation in the case of collective
redundancies. Moreover, a new permanent contract for full-time employees in small firms has been
introduced with an extended trial period of one year.
…and reduced dismissal costs for permanent workers
The labour market reform has improved the de jure flexibility of the collective bargaining system in a
way rarely found before the crisis both in Spain and in other countries of the Euro area. OECD indicators
on employment protection legislation (EPL) indicate that the reform has significantly reduced the rigidity
of the Spanish legislation on dismissals. Nevertheless, severance pay for permanent workers in Spain
remains among the highest in OECD countries, despite the significant reduction in compensation for unfair
dismissal brought about by the reform.
The reform has contributed to significant wage moderation…
The changes of internal-flexibility and collective-bargaining regulations have contributed to the
significant wage moderation observed in Spain over the past year, even if part of this moderation is the
result of protracted adverse cyclical conditions and public-sector wage cuts. While this wage moderation is
affecting workers’ living standards, there is already evidence that it has started yielding its dividends in
terms of employment performance and has contributed to save jobs. Moreover, once growth is restored, the
greater scope for firm-level collective bargaining could allow a better reflection of productivity gains into
wages. Overall, the Spanish economy appears to have made substantial progress in achieving wage
moderation, even if continuous monitoring of the effects of the reform in this area is advisable, and the
government must be ready to implement further action if performance worsens. In addition, trends in
income inequality should be monitored in order to guarantee that cost and benefits of the reform are
equally shared.
…and increased hiring on permanent contracts…
The reform has contributed to promote hiring, in particular on permanent contracts. The empirical
analysis in the report suggests that the reform could be considered responsible for about 25 000 new
permanent contracts each month, with the effect concentrated in small and medium firms (those below 100
![Page 9: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/9.jpg)
8 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
employees). The reform has also contributed to containing the duration of the unemployment spells, in
particular due to faster transitions into permanent contracts for those workers entering unemployment after
a temporary job. The analysis also shows some signs that separations decreased after the reform, especially
for temporary contracts, possibly resulting from the greater use of internal flexibility measures as an
alternative to contract termination. All these findings point to a positive effect of the reform in dampening
the widespread segmentation of the Spanish labour market, although the impact is so far small and it will
take time before the duality of the labour market is considerably reduced.
…even if, prior to the August 2013 legal change, it also led to a decline in collective dismissals
By contrast, the reform decreased collective dismissals more than any visible effect on individual
dismissals, probably due to the increased procedural uncertainty that followed the elimination of the
requirement for administrative authorisation – and the consequent increase in judicial uncertainty – as well
as the extension of the circumstances in which companies making these dismissals have to pay a tax and
carry out a special training and relocation plan. These factors are also probably one of the main drivers
behind the lack of effect of the reform on hiring on permanent contracts by large employers. In that
respect, the Government quickly reacted by addressing some of the key elements of procedural uncertainty
through a new legislative changes in August 2013. It is too early to say, however, whether these changes
will suffice to significantly reduce judicial uncertainty concerning collective dismissals in the future.
The reform has the potential to boost productivity growth and competitiveness…
Reforms of dismissal regulations, by enhancing labour reallocation, have the potential to boost
productivity growth in the long-run. Drawing from a simulation exercise based on the experience of OECD
countries over a long period of time, the report suggests that the reform could potentially boost labour
productivity growth in the long run by about a ¼ of a percentage point annually in the business sector
(excluding agriculture, mining, fuel and professional services).
…but should be accompanied by further effort to promote greater competition in product markets and
efficient activation policies
The full impact of the reform on job creation will depend however on complementary reforms in other
areas. In particular, reforms in product and service markets will not only increase competition and boost
productivity, but also improve labour market performance. The effectiveness of active labour market
policies is also very important in order to successfully place unemployed workers into jobs. The changes
introduced recently in this area go in the right direction, but a careful evaluation of these recent reforms is
necessary to ensure that the Spanish income support system to the unemployed and employment services
are able to effectively and quickly encourage and help the unemployed to find work.
While a good step, further actions could be envisaged
Overall, the 2012 labour market reform appears to have brought more dynamism into the Spanish
labour market and is a step towards the reduction of its widespread segmentation and the increase of the
competitiveness of the Spanish economy in the medium term. Nevertheless, additional adjustments could
be envisaged in order to ensure that the objectives of the labour market reform are fully attained.
In particular, the government could consider treating unlawful collective dismissals as any other type
of unfair dismissal (as in most other OECD countries), limiting the possibility of ordering reinstatement to
cases of discrimination and prohibited grounds. In addition, some of the specific, additional costs of
collective dismissals for employers introduced by the reform could be rolled back.
![Page 10: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/10.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 9
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
The evidence presented in the report also suggests that the trial period is too short for firms that are
not eligible for the contrato emprendedores. The duration of maximum trial periods for other permanent
contracts could be made longer, particularly in those cases in which this duration is currently much shorter
than in other OECD countries.
Finally, a greater convergence of employers’ costs of termination for permanent and temporary
contracts would be desirable. This will be already the case by 2015 for firms with less than 25 employees –
for fair terminations – due to a severance-pay subsidy in place for these firms and the schedule of increases
in severance pay for fixed-term contracts approved in 2010. Nevertheless, the government could consider
deepening this convergence process, notably by reducing ordinary severance costs for large employers to
align them closer to the OECD and European averages.
![Page 11: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/11.jpg)
![Page 12: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/12.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 11
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
INTRODUCTION
The global financial and economic crisis hit the Spanish labour market particularly hard. The
harmonised unemployment rate in Spain reached 26.7% in October 2013, the second highest in the OECD
area (only exceeded by Greece) and more than three times the OECD average (7.9%). Since the start of the
crisis, the number of unemployed has increased by more than 4 million persons in Spain. Even more
worryingly, long-term unemployment – defined as those unemployed for 12 months or more – as a share of
all unemployment has risen from 19.1% in in fourth quarter of 2007 to 50.4% in the third quarter of 2013.
Despite this massive increase in unemployment, the wage component of unit labour costs in the business
sector was slow to adjust. It did not decline in the first four years of the crisis, and its level at the end of
2011 was above that at the onset of the crisis.
The deterioration of the competitiveness of the Spanish economy dates back to the beginning of the
2000s, but the further decline during the early phases of the crisis and the upsurge of unemployment led to
a reaction by the Spanish government with a series of structural reforms. A first labour reform in 2010 –
among other provisions – increased severance pay for temporary contracts (fully effective by January 2015
only) and generalised the severance pay subsidy scheme to all employers and types of dismissal for
permanent contracts signed after the date of approval of the reform.1 A second major reform was
undertaken in February 2012. This reform consisted of two main elements. First, it gave priority to
collective bargaining agreements at the firm level over those established at the sector or regional level and
made it easier for firms to opt-out from a collective agreement and implement internal flexibility measures
as an alternative to job destruction. In addition, the previous practice of extending collective bargaining
agreements after their end date in the case that no new agreement could be signed by the social partners
was limited to a maximum period of one year. Second, the provisions of Spain’s Employment Protection
Legislation (EPL hereafter) were significantly modified, reshaping the definition of fair economic
dismissal, reducing monetary compensations for unfair dismissal and eliminating the requirement of
administrative authorisation for collective redundancies. In addition, a new permanent contract for full-
time employees in small firms was introduced, entailing an extended trial period of one year.
The objective of this study is to evaluate the labour market impact of the 2012 reform. In particular,
this report makes a threefold contribution. First, it places the reforms of the Spanish labour market in the
context of reforms in other OECD member countries. Second, it assesses the impact of the reform on wage
moderation and labour costs, in particular by looking at the way it appears to have shaped the evolution of
labour costs in the business sector. Finally, it evaluates the impact of the reform on labour market flows
and duality, by looking at hiring, separations, tenure on the job and duration of unemployment spells. As
far as the labour market duality is concerned, special attention is devoted to flows to and from permanent
contracts. To the extent that certain provisions included in the reform apply only to firms below certain
size thresholds (notably 50 and 25 employees), whenever possible, the effects are estimated separately for
firms above and below these thresholds.
Almost paradoxically, the comprehensive nature of the 2012 reform makes its evaluation a difficult
task. In fact, the inclusion of a large number of provisions, sometimes explicitly targeted at different
groups, does not allow the identification of a suitable control group. The reform also occurred in the
1. This provision was, however, never fully implemented and partially withdrawn in 2012 (see Section 1).
![Page 13: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/13.jpg)
12 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
middle of a double-dip recession, at a time when the economic juncture restarted to deteriorate after a short
recovery that was so hesitant that employment continued to fall. More importantly, this evaluation comes
only after a short time since the reform was implemented, and in many cases only short time series are
available.2 In addition, several amendments and implementation decrees were made at different dates after
the approval of the reform in February 2012. For all these reasons, the evaluation exercise provided in this
report should be considered as very preliminary and mainly limited to the short-time impact of the reform.
The report is organised as follows: Section 1 describes the main features of the 2012 reform. Section 2
places these features into an international perspective by comparing the Spanish institutions of collective
bargaining and employment protection legislation with those of other European and OECD countries,
documenting the country’s progress vis-à-vis those of other countries. Section 3 summarises the results of
several econometric exercises evaluating the impact of the reform on unit labour costs and various types of
worker flows, disentangling the effects by firm size whenever possible. The full details of the econometric
estimations are provided in the Annex. A simulation of the likely long-run impact of the reform on
productivity growth is also presented in this section. Section 4 briefly discusses interactions with other
policies, with particular attention to active labour market policies, stressing the potential role of other
institutions in maximising the employment dividend of the reform. The last section provides some
concluding remarks and sets forth a few recommendations for further action in this area.
1. Key components of the 2012 reform
The 2012 Spanish labour market reform was approved by the government in February 2012 as a Real
Decreto Ley 3/2012 and confirmed with no substantial modifications as the Ley 3/2012 de medidas
urgentes para la reforma del mercado laboral3 by the Spanish Parliament in July 2012. Subsequently,
several legal provisions were introduced in order to implement the reform. This comprehensive reform
modified several aspects of the Spanish labour market regulation, including collective bargaining rules and
collective and individual redundancy procedures and costs.
The objective of the reform of collective bargaining was to restore competitiveness by aligning
labour costs more closely with productivity and allow employers to exploit more easily internal flexibility
measures as an alternative to dismissals in the presence of adverse company shocks, thereby preserving
jobs and reducing employment losses in bad times (see Section 2.1 below). The reform gave priority to
collective bargaining agreements at the firm level over those at the sector or regional level, so that
collective agreements could adjust more closely to the specific needs of a firm. In addition, firms can now
opt-out more easily from a collective agreement and pursue internal flexibility measures. For example
employers can introduce unilaterally changes in working conditions (wages, working hours, work
schedules) whenever there are objective economic, technical, production or organisational reasons.4 In
addition, in the absence of an agreement with workers’ representatives, the employer willing to opt out
may now unilaterally refer the matter to arbitration by a public tripartite body (“Comisión Consultiva
Nacional de Convenios Colectivos”, CCNCC hereafter). Once settled by arbitration, this kind of dispute is
not fully reviewable by labour courts, since arbitration decisions may be challenged before a court only on
very limited grounds (see Gomez-Abelleira, 2012). Finally, collective bargaining agreements can now be
2. The analysis contained in this report is based on data that were available in early September 2013. It covers
therefore time series up to July 2013.
3. The full text of the Decreto Ley 3/2012 is available at www.boe.es/boe/dias/2012/02/11/pdfs/BOE-A-
2012-2076.pdf and the one of the Ley 3/2012 at www.boe.es/boe/dias/2012/07/07/pdfs/BOE-A-2012-
9110.pdf.
4. More precisely, this possibility existed also in the previous legislation but was extended by the 2012
reform.
![Page 14: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/14.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 13
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
prolonged for a maximum period of only one year after their end date (the so-called period of
ultra-activity), in order to provide incentives to social partners to renegotiate rapidly new agreements
adapted to any changes in economic conditions.
Substantial changes were also introduced with respect to dismissal legislation, with the objective of
making the labour market more dynamic and less segmented, thereby increasing productivity growth and
reducing the share of precarious jobs (see Section 2.2 below). The reform redefined the conditions for a
fair dismissal, improving further upon the greater clarity already introduced by the 2010 reform. While
actual or expected losses or loss of competitiveness remain fair reasons for an economic dismissal, the new
law specifies that a dismissal is always justified if the company faces a persistent decline (over three
consecutive quarters) in revenues or ordinary income.5 In addition, and perhaps more important, the firm
does not have to prove that the dismissal is essential for the future profitability of the firm. 6
Monetary compensation for unfair dismissal was reduced to 33 days’ wages per year of seniority up to
a maximum of 24 months, compared to the previous severance pay of 45 days up to a maximum of
42 months’ wages.7 At the same time, the reform removed a worker’s right to interim wages between the
effective date of dismissal and the final court ruling (except in the case when the worker is reinstated). This
made redundant the option employers often used before the reform of declaring a dismissal unfair even
before a conciliatory procedure took place and paying upfront the corresponding severance payment (the
so-called “despido exprés”, which was de facto the most commonly-used dismissal mechanism by
employers prior to the reform),8 which was therefore removed.
As regards collective dismissals, the reform eliminated the requirement of administrative
authorisation for collective redundancies while maintaining the obligation of good-faith negotiations with
unions before serving individual notice, in line with the current legislation in most OECD countries. In
addition, the new law specifies more precisely the objective reasons under which an employer can
undertake a collective redundancy. In exchange, the firm has to carry out a special training and relocation
plan for those workers that have been dismissed if the collective dismissal affects over 50 workers.
Moreover, the reform has enlarged the set of cases in which the employer must pay a tax if the collective
dismissal involves workers aged 50 years or more. In addition, in March 2013, the Real Decreto Ley
5/2013, made liable to this contribution, under certain circumstances, not only firms making profits in the
period preceding the dismissal, but also those that end up making profits in at least two of the four years
following the dismissal, thereby making this contribution dependent on future performance.9 In August
2013, further legislative changes were made in order to reduce uncertainty regarding collective dismissal
procedures. The Real Decreto Ley 11/201310 clarifies how the negotiating committee must be established
5. The new law specifies that extraordinary income should not be taken into account.
6. Case law seems to confirm that the de jure relaxation of the definition of fair economic dismissal also
holds de facto. See for example the decision of the Sala de lo Social del Tribunal Supremo dated
20 September 2013 (STS 20-9-13, Rec. 11/2013) that specifies that judges have to establish that the
economic reasons alleged by the employer are truthful and serious, but must not assess whether the
employer’s decision is an appropriate managerial decision.
7. In the case of fair dismissal, however, severance pay remains at 20 days’ wages per year of seniority up to
12 months’ wages.
8. In fact, before the reform, workers were not eligible to backpay in the case of despido exprés. This was the
only advantage of choosing this procedure for the employer.
9. The full text available at www.boe.es/boe/dias/2013/03/16/pdfs/BOE-A-2013-2874.pdf. However, this
decree also restricts the application of this tax only to those firms where the share of dismissed workers
aged 50 years or more is higher than the share of workers over 50 in the firm.
10. The full text is available at www.boe.es/boe/dias/2013/08/03/pdfs/BOE-A-2013-8556.pdf.
![Page 15: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/15.jpg)
14 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
and the documentation that the employer has to provide at the beginning of the negotiation round. Perhaps
more importantly, the same decree sets unambiguous limits to the power of courts to declare void the
collective redundancy procedure, thereby ordering reinstatement of workers (with backpay) with no
possibility for employers to opt for additional compensation in lieu of reinstatement (as in the case of
unfair individual dismissal, see above). Furthermore, it limits the cases where workers can individually
challenge collective dismissal agreements.
A new full-time permanent contract for small firms (under 50 employees) was created (Contrato de
Apoyo a Emprendedores) that allows an extended trial period of one year for firms that have not engaged
in collective or unfair dismissals in the 6 months before the starting date of the contract. This “contrato
emprendedores” includes several hiring incentives and fiscal rebates for firms under 50 employees. In
addition, the reform extended the existing subsidy equivalent to 40% of ordinary severance pay (8 days per
year of service, paid by a wage guarantee fund – FOGASA) to all cases of fair dismissal in the case of
firms with less than 25 workers.11 As the 2010 reform raised severance pay for fixed-term contracts to
12 days per year of service (from 2015), small employers will be soon liable for disbursing the same
amount for open-ended and temporary contracts, in the case of fair dismissal.
Other changes introduced by the reform include the re-instatement of the two-year maximum period
for extension of standard fixed-term contracts (which was temporarily suspended in August 2011), an
increased flexibility to use part-time contracts and the Contrato de formación y aprendizaje as well as new
regulations for training provision and the authorisation given to private temporary placement agencies to
operate in the domain of placement of unemployed workers, alongside the public employment system.
2. The labour market reform in international perspective
At first glance, the 2012 reform summarised in the previous section touches upon a number of key
aspects of the Spanish labour legislation. However, its depth and effectiveness must be rigorously assessed
on two main grounds by: i) comparing the new Spanish institutions with those prevailing in other OECD
countries (in particular, European countries); and ii) assessing the effects of the reform on labour market
performance. This section provides a comparison of the evolution of Spanish labour legislation against
those prevailing in other countries, by looking first at institutions regulating collective bargaining and
internal flexibility and then at EPL. The empirical analysis of the effect of the 2012 reform on the
performance of the labour market is then developed in the following section.
2.1. Collective bargaining
In most of continental Europe, wages are typically bargained collectively and most workers are
covered by collective agreements through administrative extension (cf. Table 1). The rationale for
collective bargaining arises from the asymmetry in contracting between individual workers and employers
regarding both access to information and bargaining power. Labour laws provide framework conditions for
collective bargaining to emerge so as to rebalance the bargaining power between employers and workers.
Hence, all else equal, compared with a situation in which only individual contracts prevail, the more
developed the collective-bargaining system is, the higher the bargaining power of workers is likely to be.
11. In principle, a transitory norm contained in the 2010 reform, had already extended this 8-day-per-year
subsidy to all firms, including even in the case of unfair dismissal (for contracts stipulated after June 2010).
However, another transitory norm of the same reform that was planned to enter into force in 2012 also
established that higher social security contributions would have been levied on contracts eligible for the
severance pay subsidy. These transitory norms were cancelled by the reform.
![Page 16: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/16.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 15
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Table 1. Structure of collective bargaining systems in Europe before the crisis
Bargaining levels and coordination in the second half of the 2000s
Central
(1) Sectoral
(2) Extension Derogations Local
(3) Change in dominant
level since 1990
Coordination
Dominant level
Cen
tral BEL *** ** *** * * ***
IRLa *** * * *** ** ***
Secto
ral
AUT *** * * ***
DEU
*** * *** ** ***
ESPa * *** *** * * *
FIN
*** ** * * 2→1,
1→2,3 *
GRCa * *** ** * * *
ITAa *** * * * ***
NLD *** *** ** * ***
NOR * *** * * * ***
PRTa *** *** * *
Lo
ca
l
CZE ** ** *** *
DNK
**
*** 2→3 **
FRAa * ** *** * *** 2→3 *
GBR * ***
HUN * * * * *** *
POL ** * * ***
SVK ** * *** 1,2→3 *
SWE ** *** 2→3 **
Notes: 1 refers to central level of bargaining, 2 to sectoral and 3 to local. * = low ; ** = medium ; *** = high, qualifying the relative importance for bargaining levels and the importance of coordination. The table should be read by line, as it describes the relative importance of the various bargaining levels and of the extension of, and derogation from, sectoral agreements within each country. It is not meant to provide an assessment of the relative importance of a given bargaining level across countries.
a) Collective bargaining systems incurred significant changes in Greece, France, Ireland, Italy, Portugal and Spain after the start of the global financial crisis; they are not included here as the available information is not systematic for all other countries.
Source: OECD (2012), Du Caju et al. (2008), van Klaveren (2011).
From a theoretical point of view, the level at which collective bargaining takes place can yield
substantially different outcomes. In general, aggregate flexibility brought about by centralised or
coordinated bargaining allows wage-setters to internalise externalities associated with wage increases and
delivers better outcomes in terms of unemployment, inflation and innovation by incumbent firms.
Conversely, relative flexibility associated with firm-level bargaining typically spurs innovation by entrants
and allows a better adjustment of wage growth to firm-level productivity growth, thereby allowing to save
jobs in the presence of idiosyncratic shocks. By contrast, in the case of bargaining at the sectoral or
regional level, cross-sector imitation often pushes wages upwards in times of boom and delays the required
wage-adjustments in times of crisis, particularly in the presence of heterogeneous firm performances (see
e.g. OECD, 1994, 2006, Flanagan, 1999, Haucap and Wey, 2004, Jimeno and Thomas, 2013).
In the 1990s, in a number of European countries where sectoral bargaining was playing a major role,
governments pressed for national-level agreements. In this case, the key objective was often to limit the
automatic indexation of wages to inflation. In fact, monetary integration ruled out external adjustment via the
![Page 17: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/17.jpg)
16 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
exchange rate to compensate for losses in national competitiveness, and, starting from 1999 for EMU
countries, the use of monetary policy instruments to adjust to asymmetric shocks. As a result, the burden of
adjustment to economic imbalances and shocks shifted increasingly onto the labour market. The outcome of
this process was that, in most countries of the Eurozone for which data are available (see Figure 1), less than
50% of companies had an automatic indexation of wages at the onset of the crisis – and, in many countries,
automatic indexation was not imposed by collective agreements but the result of independent company
policies. In that respect, Spain was among the countries where indexation was more frequent – covering
about 70% of firms – which contributed to the erosion of Spanish competitiveness before the crisis, as shown
by the pre-crisis evolution of its unit labour costs in comparison with other Euro countries (see Figure 2).12
Figure 1. Percentage of firms subject to a mechanism of adjusting base wages to inflation before the crisis
Selected Euro countries
Note: Countries are selected on the basis of available data. Data refer to 2007-2008
Source: Eurosystem’s WDN Survey (Druant et al., 2009, European Central Bank, 2010).
Figure 2. Evolution of unit labour costs in Euro countries, 2000Q1-2008Q1
Percentage change
Note: Only countries that joined the Euro before 2007 are included. Data for Greece are not available.
Source: Eurostat.
At the same time, in most OECD countries, the role played by collective negotiations at the
firm/establishment level has increased, leading to a significant decentralisation of collective bargaining
12. However, other structural factors – such as the dynamics of the sectoral composition of the economy – are
also responsible for this trend in unit labour costs.
0
20
40
60
80
100
120
ITA AUT IRL FRA GRC PRT ESP BEL LUX
-5
0
5
10
15
20
25
30
35
40
Germany Austria Euro area Finland Belgium Netherlands France Italy Portugal Luxembourg Spain Ireland
![Page 18: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/18.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 17
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
systems since 1990. While this process often started in the 1980s for bargaining over working-time
reduction, it has since extended to matters of pay. Decentralisation has taken place in two main ways:
i) instead of setting standard wage increases, sectoral agreements increasingly provide a framework for
firm-level agreements; and ii) sectoral agreements increasingly include derogation clauses, allowing
firm-level agreements to depart from sectoral agreements in specified cases.13
The substance of sectoral agreements has been changing in many countries, leaving much more room
for firm-level bargaining on wages. In some cases, the determination of average wage increases still takes
place at the sectoral level, but the decision on how to distribute wage increases among employees is left to
firm-level bargaining.14 However, at least before the onset of the crisis, in most countries, including Spain,
firms covered by a multi-employer agreement were bound to observe what is called the “favourability
principle”, meaning that firm level agreements had to be more favourable to employees than higher-level
agreements. As a result, firm-level bargaining resulted often in higher wages with respect to minima
bargained at the sector level. For example, exploiting transitions from one regime to another, wages
negotiated at the firm level are found to be higher in Denmark than those negotiated under sectoral
agreements, even though significantly more dispersed (Dahl et al., 2011). Similar results are found by
Gürtzgen (2010) and Rusinek and Rycx (2013) for Germany and Belgium.15
The inclusion of derogation clauses in sectoral, regional or national collective agreements, allowing
firms to opt-out of higher-order agreements, has become more frequent, especially in Germany and Ireland.
From a theoretical viewpoint, Jimeno and Thomas (2013) show that sectoral bargaining systems can deliver
similar labour market performances as decentralised systems, if firms and workers are not prevented from
agreeing to opt out of higher-level agreements. Derogation clauses are typically of two types: “hardship” or
“inability-to-pay” clauses, which allow temporary deviations from higher-level agreements for firms facing
economic difficulties; and “opening” or “opt-out” clauses, which can be invoked either by firms that cannot
afford to meet the general standard, especially small and medium-sized enterprises, or by firms facing threats
to future competitiveness and possible relocations of investment and production sites (Visser, 2004). In most
countries, the use of derogation clauses remained limited before the crisis (see Table 1), with few exceptions.
The “inability-to-pay” clause included in the central agreements since 2003 was regularly used in Ireland
(van Klaveren, 2011). But, of particular note, actual use of these clauses expanded significantly in
Germany.16 In this country, industries covered by sectoral agreements allowing derogation clauses
experienced greater net employment growth and lower job destruction (Brändle and Heinbach, 2010). This
trend can be considered one of the most important factors that restrained the growth of unit labour costs in
Germany before the crisis (see Figure 3).17
13. In addition, the share of variable pay, by definition negotiated at the firm level, has been growing.
14. This is a very common practice in Denmark and Sweden, and exists in Austria, Belgium, Germany and
Italy. In the Netherlands, a large share of employees are covered by sectoral agreements which allow for
choices between pay and working time to be decided at the firm level. In the Czech Republic, Denmark,
and the Slovak Republic, industry-level agreements increasingly tend to set minima while actual wage
increases for the rest of the pay scale are negotiated at the firm level (see OECD, 2012).
15. In Germany, however, administrative extension of collective agreements is limited and occurs almost on a
voluntary basis, except when automatic extension to all workers and companies of an industry is imposed
by the Federal Ministry of Labour. Many employers adhere to the collective wage agreement concluded for
their respective industry even if they are not member of a business association that signed it. Nonetheless,
they have no obligation to do so. Nevertheless, new companies tend to resist applying branch-level
collective agreements and, as a result, the proportion of workers covered by collective agreements has gone
down over time, particularly in Eastern Germany (Düll, 2013).
16. Increased use of derogation clauses resulted in cuts in basic pay, reductions in agreed wage increases, lower
wage rates for job starters or reduction/suspension of bonuses (Keune, 2010; Haipeter and Lehndorff, 2009).
17. Other factors to be mentioned, however, are the expansion of low-pay jobs and the reduction of workers
covered by collective agreements
![Page 19: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/19.jpg)
18 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Even if derogation clauses were possible in Spain before the 2012 reform, they were seldom applied.
Overall, the Spanish economy was typically characterised by a much stronger reliance on employment
adjustments to absorb shocks. For example, before the onset of the crisis, the main adjustment strategy to a
demand shock for about 70% of Spanish firms was to reduce employment– mainly by suppressing temporary
jobs – while this strategy was preferred by only 40% of firms, on average, in other countries according to the
Eurosystem’s WDN Survey (Figure 3, Panel A). In fact, wage cuts or wage freeze were very rarely
undertaken by Spanish firms in the five years preceding the crisis in comparison with other countries (Figure
3, Panel B). Evidence from the follow-up WDN survey, conducted in the first months of the crisis, shows
that, wage cuts and wage freezes remained much less frequent in Spain than in most other European
countries, despite the severity of the 2008-2009 recession in the country (European Central Bank, 2010).18
Figure 3. Adjustment strategies to adverse shocks used by European firms before the crisis
Panel A. Percentage of firms for which job destruction is the main adjustment strategy to an adverse demand shock, 2007-2008
Panel B. Percentage of firms having frozen or cut wages between 2003 and 2008
Note: Countries are selected on the basis of available data
Source: Eurosystem’s WDN Survey (Fabiani et al., 2010, Babecký et al., 2009, European Central Bank, 2010).
18. Some caution is required in the analysis of the follow-up WDN survey since it is smaller and with a larger
non-response rate than the original survey.
0
10
20
30
40
50
60
70
80
HUN AUT PRT POL SVN IRL EST LUX NLD FRA ITA CZE BEL ESP
Mainly temporary jobs Mainly permanent jobs
0
5
10
15
20
25
30
35
40
ESP ITA SVN HUN FRA IRL GRC POL LUX BEL PRT AUT NLD EST CZE
Wage cut Wage freeze
![Page 20: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/20.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 19
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
In this context, the 2012 reform appears to have improved the flexibility of the Spanish collective
bargaining system. In particular, the introduction of the principle of dominance of firm-level agreements
on higher-level ones and the greater possibility of opting-out of collective agreements even in the absence
of consensus among social partners at the company level represent potential instruments of internal
flexibility that, as shown above, were hardly found both in Spain and the other European countries before
the crisis. Whether or not there is evidence that these instruments are contributing to promote the
competitiveness of Spanish firms is a question that will be examined in Section 3.
2.2. Employment Protection Legislation
There is a large body of theoretical and empirical research – summarised in OECD (2013a) among
others – that suggests that hiring and firing regulations have a significant impact on the reallocation of
labour resources, while the long-run impact on aggregate unemployment rates is, at best, small. Countries
with relatively high employment protection are typically found to have lower hiring and separation rates
and a slower resilience to output shocks. Moreover, there is evidence that in these countries stringent
regulations stifle the allocation of labour to the most productive uses, thereby hindering productivity and
economic growth. Perhaps more important in the Spanish case, if stringent employment protection on
regular, open-ended contracts coexists with relatively easy access to fixed-term contracts, firms tend to
react by substituting fixed-term for permanent contracts – due to the smaller cost involved with the
termination of the employment relationship at the end of a fixed-term contract (see e.g. Boeri and
Garibaldi, 2007; Bentolila et al., 2008, 2012; Jahn et al., 2012) – with no significant impact on employment
and negative effects on productivity growth, in particular if the likelihood of contract conversion is small
(see e.g. Dolado et al., 2012, Cappellari et al., 2012).
OECD indicators on EPL allow benchmarking cross-country differences in the stringency of hiring
and firing regulations. These indicators quantify, for employers of large companies, the costs and
procedures involved in dismissing individuals – or groups of employees – or hiring workers on fixed-term
or temporary-work-agency contracts, as in force on the 1st of January of each year. The indicators
concerning the regulations governing individual and collective dismissals of workers with regular,
open-ended contracts, are perhaps the most useful for policy analysis. These indicators cover: i) procedural
inconveniences that employers face when starting the dismissal process, such as notification and
consultation requirements; ii) notice periods and severance pay, which typically vary by tenure of the
employee; iii) difficulty of dismissal, as determined by the circumstances in which it is possible to fairly
dismiss workers, as well as the repercussions for the employer if a dismissal is found to be unfair (such as
compensation and reinstatement); and iv) additional restrictions for collective dismissals, such as
additional delays, costs or notification procedures when an employer dismisses a large number of workers
at one time. Although any quantification is inevitably a simplification (see e.g. OECD, 2013), and users
need to be cautious in using these indicators to assess the relative position of a given country, the change in
value of the summary indicator for regular contracts for Spain between 2008 and 2013 suggests that the
recent reforms significantly reduced the rigidity of the Spanish legislation on dismissals (Figure 4).
The OECD indicators can also be used for a rough comparison, component by component, of the
regulation in Spain with respect to other OECD countries. The Spanish economy appears now more
flexible than the average OECD country in areas such as notification procedures and the length of notice
periods (see Figure 5). In particular, notice periods were reduced in 2010 from 1 month to two weeks,
independently from the length of service, which makes of Spain one of the OECD countries with the
shortest notice periods, particularly for workers with a long tenure. Indeed, average notice periods in
OECD countries are 3.5 weeks, 1.3 months and 2.7 months for workers with, respectively, 9 months,
4 years and 20 years of tenure at the time of dismissal (OECD, 2013b).
![Page 21: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/21.jpg)
20 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Figure 4. Protection of permanent workers against individual and collective dismissal, 2008 and 2013
Note: The indicator vary from 0 (least restrictive) to 6 (most restrictive).
Source: OECD Employment Protection Database, 2013 update.
Figure 5. Employment protection for permanent workers: Spain vs. OECD
Disaggregate components, 2008 and 2013
Note: Indicators vary from 0 (least restrictive) to 6 (most restrictive) and refer to rules applicable to large companies.
Source: OECD Employment Protection Database, 2013 update.
0
1
2
3
4
5
6
2013 2008 2013 OECD average
0
1
2
3
4
5
6
7
OECD 2013 ESP 2013 OECD 2008 ESP 2008
![Page 22: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/22.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 21
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
In contrast with a number of countries in continental Europe, Spanish workers have no right to
reinstatement in the case of unfair individual dismissal19 – except when dismissal is based on prohibited
grounds (such as discrimination) – which provides greater certainty to employers about the cost of
dismissal. This is somewhat counterbalanced by a level of monetary compensation for dismissal which is
high in cross-country comparable terms. In fact, as far as large companies are concerned, Spain remains
among the countries with the most expensive requirements as regards severance pay both in the case of fair
and unfair dismissal. And this occurs despite the significant reduction in compensation for unfair dismissal
brought about by the 2012 reform (see Section 1), which put additional compensation in this case more in
line with the amounts paid in other countries (see Figure 5). More precisely, in the case of unfair dismissal,
the typical monetary compensation in OECD countries is about 13.7 monthly wages at 20 years of tenure
(against about 22 months in Spain after the reform). For the same length of service, ordinary severance pay
in the case of fair dismissal is on average 6 monthly wages among OECD countries where dismissed
employees are entitled to it, while it is equal to 10 monthly wages in Spain, in the case of companies with
25 employees or more. Finally, Spain is one of the only 7 OECD countries where workers in large
companies are entitled to ordinary severance pay even if their job tenure is shorter than 1 year (see OECD,
2013b).
The situation is comparatively better for small firms, and improved further after the 2012 reform,
insofar as the latter included several provisions targeted at them. In the context of the Spanish economy,
where small firms represent a large share of total employment, this is an important potential source of
dynamism for the labour market.20 In particular, the severance-pay subsidy covered by the wage guarantee
fund, which results in a 40% reduction of the burden of ordinary severance pay borne by employers with
less than 25 employees, is applicable to all cases of fair dismissals after the 2012 reform. This policy action
has put the cost of ordinary severance payment at the level of the OECD average for firms of that size –
excluding those countries with no compulsory severance payments in the case of fair dismissal.
The case of the trial period is somewhat similar. For ordinary permanent contracts the maximum
length of the trial period is on average shorter – and therefore more constraining – than in most other
OECD countries (4 months in Spain against an OECD average of 5.1 months), and can be as short as
two weeks in certain sector-level collective agreements. Nonetheless, the introduction of the “contrato
emprendedores” (see Section 1) has de facto made the trial period for firms with less than 50 employees
the longest within civil-law OECD countries.21 This is particularly important since, although the reform has
significantly relaxed the definition of fair dismissal for economic reasons, dismissal for personal reasons
remains relatively difficult in Spain. In principle, therefore, the longer trial period should allow small
employers to better experiment and screen new employees, reducing the risk of bad matches and thereby
increasing hiring incentives. Indeed, there is evidence in the literature that the length of the trial period and
the extent of reinstatement are usually the two EPL components with the greatest impact on labour
reallocation (see Bassanini and Garnero, 2013). Yet, a very recent decision of a local court stated that, for
certain occupations, a trial period as long as one year would be in violation of article 4.4 of the European
19. However, reinstatement with backpay can be ordered when a collective dismissal is declared void by courts
(see below).
20. According to the OECD Structural Business Statistics, firms with less than 20 employees and less than
50 employees accounted for 50% and 62%, respectively, of the Spanish non-financial business-sector
employment in 2010.
21. Legal systems can be roughly divided in civil-law and common-law countries, the latter including
Australia, Canada, Ireland, Israel, New Zealand, the United Kingdom and the United States. Common-law
countries typically have lighter statutory regulations limiting the freedom of contracting parties. Beyond
most US states and Canadian provinces, where claims for unfair dismissal cannot be filed, the United
Kingdom has the longest trial period (2 years for all firms), followed by Ireland and Israel (1 year for all
firms) and Australia (1 year for firms with less than 15 employees; see OECD, 2013b).
![Page 23: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/23.jpg)
22 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Social Charter, which stipulates the right to advance notice for all dismissed workers.22 However, it is too
early to say whether this argument will be incorporated into other court decisions in the future, and the
consequences that this will have for the application of the reform. Nonetheless, to avoid these types of
conflict with supranational legislation, a number of OECD countries have introduced the obligation of
advance notice of dismissal even for workers in the last months of the trial period, when the latter is long.23
By contrast, before the reform, Spain was one of the few OECD countries where a specific
administrative authorisation was required for collective dismissals. At the beginning of 2012, a similar
institutional arrangement could be found only in Greece and Mexico where an administrative authorisation
is required if no agreement is reached among social partners on the terms of the collective dismissal
procedure. Yet, the administrative authorisation may help reducing uncertainty about the final cost for
employers of dismissal decisions in countries where judicial review of these decisions is extensive and
employers’ choices can be nullified by the court. For example, in May 2013, France re-introduced the
requirement of administrative authorisation in the absence of agreement among social partners in order to
reduce the degree of judicial uncertainty generated by the possibility for courts of invalidating social plans
associated with collective dismissals and, consequently, ordering the reinstatement of involved workers,
sometimes several months after dismissal.
There is some evidence suggesting that the Spanish experience immediately after the 2012 reform
closely resembles the French experience before the 2013 reform. Even if the litigation rate as regards
collective redundancies appears to have remained low (below 5% according to estimates by Palomo-Balda,
2013), a large share of concluded proceedings concerning collective dismissals resulted in court rulings
against the employer. And, in most of these cases, the judges ruled that the dismissal procedure was nul
and void and ordered the reinstatement of the affected workers with backpay – a novel situation in the
Spanish labour market that was essentially inexistent before the 2012 reform (see Palomo-Balda 2013,
Mercader-Uguina and de La Puebla-Pinilla, 2013, and Ministerio de Empleo y Seguridad Social, 2013).24
In addition, in most cases, the court decisions against employers were not based on challenges to the
alleged substantive reasons for dismissal but on the non-respect of the negotiation procedure, in particular
as regards the determination of employees’ representatives in the consultation committee, the
incompleteness of the documentation provided by the employer at the start of the consultation process, or
the lack of good faith in consultations. Given these outcomes, the low rate of litigation might not be fully
22. Juzgado de lo Social 2 de Barcelona (Sentencia 412/13, 19-11-2013).
23. This is the case, for example, in Germany, Ireland, Switzerland and the United Kingdom. Often these
notice periods are, however, shorter than in the case of other permanent workers. Probationary periods,
when too long, could also conflict with the requirements of the ILO Convention 158 for those countries
which have ratified this convention (including Spain). Article 2 of this convention stipulates that
probationary periods should be of reasonable duration. On this basis, in France (one country that signed
Convention 158) the Contrat Nouvelle Embauche that was introduced in 2005 and allowed a 2-year trial
period under certain conditions was subsequently annulled by the administrative high court because it was
considered in violation of the ILO convention (see e.g. Lepage-Saucier et al., 2013). However, other
countries that signed Convention 158 have long probationary periods. In Australia, for example, they are as
long as one year, for small firms, and six months, for large firms, suggesting that these durations can well
be considered as “reasonable”, at least conditional on firm size, even as regards the application of the ILO
convention (see also the observations on Australia of the ILO Committee of Experts on the Application of
Conventions and Recommendations, 2012). At the time of writing, Spanish labour unions have filed a
complaint with the constitutional court concerning the possibility that the contrato emprendedores is in
violation of Convention 158.
24. Before the reform, a judgement of nullity of a collective dismissal procedure that had been previously
authorised by the labour authority was basically impossible without proving an explicit misbehaviour of
the labour authority (see e.g. Mercader-Uguina and de La Puebla-Pinilla, 2013).
![Page 24: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/24.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 23
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
representative of the effective cost for employers, who might be either forced to give concessions during
the negotiation phase or tempted to chain together a smaller number of individual dismissals over
subsequent periods of three months in order to escape from collective dismissal regulations,25 even though
these chained series of dismissals would be in principle illegal if motivated by the same economic cause.
These shortcomings of the new regulations of collective dismissals have been recently addressed by
the government with the approval of the Real Decreto Ley 11/2013 in August 2013, which has more
explicitly defined the requirements of the consultation procedure and the cases in which the dismissal can
be declared void. In addition, the interpretation of the law has been further clarified by recent court rulings
of the Supreme Court, specifying in particular that the absence of an agreement with unions do not justify
per se the invalidation of the dismissal procedure and defining how extensively the consolidated accounts
of a business group must be considered in assessing the justification of the dismissal decision.26 It is too
early to say, however, whether these interventions will prove sufficient to restore legal certainty as regards
to the cost and outcomes of collective dismissals. Yet, albeit significantly reduced, the discretionary role of
courts to invalidate a collective-dismissal procedure and order reinstatement remains substantial, at least on
paper.
However, the reform has also increased certain, specific costs of collective dismissal by making
compulsory to set a retraining and redeployment plan – in the case of dismissals involving more than
50 employees – and enlarging the set of circumstances in which a contribution to the treasury must be
made (see Section 1). De la Puebla-Pinilla (2013) estimates these costs as substantial: the total cost of
additional taxes and retraining/redeploying affected workers can be greater than the total amount of
severance payments that the company would have disbursed had it been able to unfairly dismiss these
workers.27 These additional provisions have not been modified by the August 2013 revision of the reform.
Finally, the Spanish labour market is characterised by a high level of duality. At the onset of the crisis,
31.7% of wage and salary employment was on a fixed-term contract (Figure 6). This proportion went down
significantly during the crisis, because of the extremely high rate of temporary job destruction, but remains
among the highest in the OECD. The duality of the Spanish labour market emerges even more clearly from
data on new hires. In 2007, 88.3% of new labour contracts were fixed-term, and this figure increased
during the crisis peaking at 92.3% in 2011, to decline only slightly afterwards.28 Although the share of
25. See for example “Los trabajadores de Pastas Gallo denuncian un ERE encubierto”, El País, 13 September
2013. Collective dismissal regulations must be applied if, within 90 days, the employer plans to dismiss:
more than 10 workers in firms with less 100 employees; at least 10% of workers in firms with at least 100
and less than 300 employees; and 30 workers in firms with at least 300 employees.
26. See in particular the decisions of the Sala de lo Social del Tribunal Supremo dated 27 May 2013 (STS 27-
5-13, Rec. 78/2012), 20 September 2013 (STS 20-9-13, Rec. 11/2013) and 25 September 2013 (STS 25-9-
13, Rec. 3/2013). Moreover, future decisions of the Supreme Court are likely to build up a coherent body
of case law that will help reducing further legal uncertainty. In fact, the 2012 reform, by allowing
complaints concerning collective dismissals to be filed directly with regional labour courts, makes it
possible for the Supreme Court to review all collective dismissal cases, as long as the party who has lost in
the first instance lodges an appeal (Gomez-Abelleira, 2012).
27. According to Spanish legislation, if the number of dismissed workers within a period of 90 days is larger
than the threshold triggering collective dismissal procedures, the dismissal is void if the firm does not
implement the procedure at all. This implies that unfair dismissal (“despido improcedente”) is not an
option open to employers in this case. As discussed above, however, in certain cases companies might
decide to limit the number of affected workers, and/or chain individual dismissals over a long-period to
avoid paying the higher costs associated with collective dismissals, even at the risk of having the individual
terminations be ruled as unfair by a court.
28. Data on new contracts are from Datos Estadísticos de Contratos, published by the Servicio Publico de
Empleo Estadal (SEPE).
![Page 25: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/25.jpg)
24 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
fixed-term contracts in new hires is typically high in countries with stringent EPL, the Spanish labour
market appears to perform particularly badly in this respect. For example, in both France and Italy the
share of fixed-term contracts in new hires was much lower in 2011 (78% in both countries; see Paraire,
2012, and Ministero del Lavoro e delle Politiche Sociali, 2012).
Despite the high share of fixed-term contracts among new hires, the rate of contract conversion is
extremely low in Spain. Only 3% of all new contracts are converted from fixed-term into permanent
contracts. From a worker perspective this implies that transitions across temporary and permanents statuses
are rare (e.g. Güell and Petrongolo, 2007). Even taking a longer horizon, the evidence is qualitatively
similar: about one-third of the workforce that is found in temporary jobs at a certain point in time
permanently rotate across temporary jobs and unemployment for at least 6 to 8 years, and even when they
obtain an open-ended contract it is unstable (Garcia-Serrano and Malo, 2013). In other words, the Spanish
labour market is dynamically segmented between well-protected “insiders” and precarious “outsiders”,
who cycle between temporary jobs and unemployment with little hope of transiting towards permanent
jobs.
In contrast with the legislation of most other countries, severance pay is due in Spain in the case of
termination of a temporary contract at the initiative of the employer even at the end date, but its amount is
small.29 In 2010, the government approved a small reform of these payments, entailing their increase from
8 to 12 days per year of service by 2015. While for firms with less than 25 employees this will equalise the
cost of fair termination between temporary and permanent contracts, this cost will remain nonetheless
much lower for temporary contracts in the case of large firms – and for all firms in the case of unfair
dismissal.
Figure 6. Incidence of temporary employment, 2007 and 2012
Employees with temporary contracts as a percentage of total employees
Source: OECD Database on Labour Force Statistics.
29. France, Israel and the United Kingdom are among the few OECD countries where temporary workers have
a statutory right to severance pay even in the case of fair termination.
0
5
10
15
20
25
30
35
EST AUS GBR SVK LUX BEL NOR DNK CZE AUT HUN GRC IRL TUR CHE ISL CAN JPN ITA DEU FRA FIN SVN NLD PRT ESP POL CHL
2012 2007
OECD 2012 average
![Page 26: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/26.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 25
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
3. The labour market impact of the 2012 reform
Almost paradoxically, the comprehensive nature of the 2012 labour market reform makes its evaluation
a difficult task. In fact, the inclusion of a large number of provisions, sometimes explicitly targeted at
different groups, does not allow the identification of a well-defined control group. The reform also occurred
in the middle of a double-dip recession, at a time when the economic juncture started to deteriorate again
after a short recovery that was so hesitant that employment had continued to decline (see e.g. Ministerio de
Empleo y Seguridad Social, 2013). This suggests that one of the few available methods to analyse the role of
the reform in the recent evolution of the Spanish labour market is through regression-discontinuity models in
which the effect of the reform is identified through discontinuous patterns occurring at the time of its
enforcement and the business-cycle is modelled through observable controls and non-linear time trends (see
Box 1). The fact that discontinuous patterns of labour market performance occurring around February 2012
are used to identify the effect of the reform makes it impossible to distinguish its impact from that of other
institutional changes occurring around the same date. Obviously, the analysis developed in this section
assesses only the joint effect of the reform and other simultaneous institutional changes.
Box 1. Estimating the impact of the 2012 reform using regression-discontinuity models
The estimation strategy followed in this report identifies the joint effect of all the provisions included in the reform by comparing labour market performance before and after February 2012. The key identification assumption is that, conditional on control variables included in the model, labour market performance evolves in a relatively smooth way, so that any discontinuous jump in performance can be attributed to the labour market reform (and other institutional changes occurring simultaneously, such as the Agreement on Employment and Collective Bargaining 2012-2014, signed by the main business associations and trade unions – CEOE, CEPYME, CCOO and UGT – in January 2012). In order to properly isolate the effect of the reform from that of the business-cycle (which is key for the validity of the smoothness assumption), the estimation models capture economic fluctuations by controlling for the standardised unemployment rate, changes in registered employment (at the regional level when microdata are used) and, most importantly, polynomial time trends up to the 5
th order. Baseline specifications include a polynomial trend up to the
highest order n such that the n-th term is significant. However, except when specified otherwise, results are robust both to changes in the order of the polynomial and to its exclusion from the specification. When a sufficiently long number of periods is available after the reform, polynomial trends are alternatively included as either homogeneous (same parameters before and after the reform) or heterogeneous (different parameters before and after the reform), and the robustness of the results to these different specification choices is assessed. The results are also robust to the use of other controls for the business cycle such as changes in industry-level value added or productivity growth or the FEDEA index of economic activity (see the Annex for details). Finally, other variables capturing compositional effects (region, industry, workforce composition, etc.) are included in each specification. As suggested by Card and Lee (2008), since the identification is based on a time discontinuity, standard errors are always adjusted for clustering at least on time (and, where relevant, also to other dimensions).
Obviously, misspecification of the empirical model might lead to identify a discontinuous shift in performance around the date of a reform even when this shift occurs before the reform (and cannot therefore be attributed to it). To validate the empirical model, therefore, placebo tests need to be run. These tests consist in setting in the empirical model an earlier date than the actual reform date to estimate possible discontinuities. If discontinuous shifts in performance are really induced by the reform, then no effect should be found at these anticipated dates. This is indeed the case for all the results discussed in this report, where placebo tests are run by anticipating the date of the reform by up to 9 months.
A second issue concerns possible manipulations around the threshold. For example, if the introduction of the contrato emprendedores were anticipated, employers eligible for the subsidy could delay hiring to after the reform in order to enjoy the subsidy. However, the details and the breadth of the reform were never mentioned in the programme of the PPE before the November 2011 elections that the party won and was not made public until well after the inaugural address of Prime Minister Rajoy made in front of the parliament at the end of December 2011. It is therefore reasonable to assume that if threshold manipulation occurred, that is if firms postponed certain choices until the approval of the reform, this phenomenon concerned, at worst, only the period January-March 2012. Under these assumptions, threshold manipulation is not an issue in the regressions estimated on quarterly data because the outcome of any firm choice, which was delayed from January to March 2012, is included in the same quarter, with no effect on the average outcome of the quarter. In the case of estimates based on monthly data, baseline models are re-estimated excluding the period from January to March 2012 from the sample as an additional robustness check.
![Page 27: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/27.jpg)
26 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
A number of provisions, however, also either entered into force, de jure or de facto, after the date of
approval of the main reform (February 2012) or were modified since then. This is the case, for example, of
the norms regulating the CCNCC (Real Decreto, 1362/2012, approved in September 2012), of those
regulating collective dismissal procedures and associated financial contributions for firms (Real Decretos
1483/2012 and 1484/2012, approved in October 2012), of the entry into force of binding time limits for the
automatic extension of expired collective bargaining agreements (July 2013) and of the corrections made
to collective-dismissal regulations in March and August 2013 (Real Decretos Ley 5/2013 and 11/2013; see
Section 1 above). This makes more difficult to identify unambiguously discontinuous patterns in the data.
Last but not least, the assessment presented here is performed shortly after the reform. Available data
cover at best 18 months in the post-reform period, but sometimes much less. Thus, estimates presented in
this section measure only the short-term impact of the reform. In addition, the effect of structural reforms
often takes time to materialise since economic agents adapt slowly to new conditions. This is likely to be
particularly the case in Spain due to the role of the courts in applying the new legislation and the time that
it will take to the judicial system to develop a uniform body of case law. All of these factors suggest great
caution in interpreting the results, which at best can be interpreted as preliminary and only roughly
indicative of a general trend.
3.1. Labour costs
As discussed in the previous sections, the objective of the reform of collective bargaining and internal
flexibility was to restore the competitiveness of Spanish firms, by ensuring a better adaptation of wages to
productivity developments. The available, preliminary evidence concerning new collective bargaining
agreements suggests that wage increases with respect to observed and expected inflation have been
restrained in 2012 and 2013 (Ministerio de Empleo y Seguridad Social, 2013, BBVA, 2013a). In addition,
even though the share of workers covered by automatic indexation of wages has not fallen significantly,
the impact of indexation clauses on wages appears to have become smaller, probably due to a more flexible
application of these clauses in the new contracts, following the recommendations of the 2012 general
agreement among social partners (Izquierdo et al., 2013).
Spain has been one of the European countries where unit labour costs have declined the most since
2012 (Figure 7). Between the fourth quarter of 2011 and the second quarter of 2013, unit labour costs
declined by 3.9%. This has occurred despite the withdrawal of most hiring subsidies decided by the
government in July 2012, which undoubtedly pushed labour costs upwards (see Ministerio de Empleo y
Seguridad Social, 2013). In the same period, in most European countries, labour costs were either on a
moderate rise or decreasing only slightly.30 Even in Portugal, which has also been hit hard by the crisis, the
drop in unit labour costs was 3 percentage points smaller than in Spain.
One needs to be cautious, however, in attributing these developments to the 2012 labour market
reform. In particular, as argued elsewhere (BBVA, 2013b), wage moderation was particularly important in
the public sector in 2012 for reasons that have nothing to do with the labour market reform, albeit closely
related to the the government’s action to fiscal consolidation.31 To shed light on this issue, therefore,
Figure 8 presents the evolution of unit labour costs separately for the business sector and non-market
services including the public administration. While the effect of the drop in civil servant’s wages is clearly
visible in the data for the fourth quarter of 2012, significant wage moderation is also observable outside the
30. Data for Greece are not available. The only other European country with a significant drop in unit labour
costs is Cyprus (-7.3%).
31. In Spain most civil servants have two bonus payments, equivalent to about one month's salary, paid twice a
year, at Christmas and in July. The Christmas bonus was withdrawn in 2012, thereby resulting in a
significant fall in 2012 public sector wages, concentrated in the fourth quarter.
![Page 28: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/28.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 27
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
public sector. In fact, in the non-agricultural business sector, unit labour costs, excluding non-wage costs,
decreased by 3.2% between the fourth quarter of 2011 and the second quarter of 2013, that is an average
annual drop of about 2.1%. By contrast, the growth of these costs was substantially flat in the three years
preceding the reform.
Figure 7. Evolution of unit labour costs in selected European countries, 2008-2013
Q4-2011 = 100
Note: Data are seasonally adjusted and adjusted for working days.
Source: Eurostat.
Figure 8. Evolution of the wage and salary component of unit labour costs in Spain, by industry, 2008-2013
Q4-2011 = 100
Note: Data are seasonally adjusted and adjusted for working days. The business-sector excludes agriculture.
Source: OECD calculation on the basis of the Spanish Quarterly Labour Cost Survey and Quarterly National Accounts.
90
95
100
105
110
115
120
125
Germany Ireland Spa in France Ita ly Portugal
80
85
90
95
100
105
110
Business sector Total economy (excluding households) Public administration and non-market services
![Page 29: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/29.jpg)
28 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Again, the evidence presented in Figure 8 is not sufficient to quantify the effect of the labour market
reform on labour cost competitiveness of Spanish firms. The prolonged recession has undoubtedly affected
the evolution of wages, and its effects are indeed visible in terms of wage moderation before the reform
(see Figure 6 and 7). However, the econometric models estimated for this report using industry-level data
allow for netting out the effect of the business-cycle, inflation and productivity growth on the dynamics of
labour costs per hour worked so that the effect of the 2012 labour market reform (and the 2012 national
agreement among social partners) can be identified through the discrete shift in the growth of labour costs
observable since the beginning of 2012 (see Box 1). The estimated results suggest that the 2012 reform
(together with the 2012 collective agreement) induced a drop in the growth of unit labour costs in the
business sector of between 1.2% and 1.9% (Figure 9).32 It can be concluded, therefore, that more than 50%
of the observed drop in unit labour can be attributed to the 2012 reform.
Figure 9. Estimated impact of the 2012 reform on the year-on-year growth of business-sector labour costs
Minimum and maximum estimated percentage-point effects
Note: Estimates obtained using different specifications of a model in which the year-on-year the industry-level harmonized index of wage and salaries per hour worked is regressed on the unemployment rate, a polynomial trend (industry-specific in certain specifications) and year-on-year changes in industry workforce composition, the consumer price index and either log employment or log labour productivity. Significance levels are obtained by adjusting errors for clustering by industry and time. The estimates are based on industry data disaggregated at the 1-digit level. **: significant at the 5% level.
Source: OECD estimation on the basis of the Spanish Quarterly Labour Cost Survey, Quarterly National Accounts and the Encuesta de Población Activa. See Annex for the detailed estimation method and results.
These results are consistent with those of a recent Bank of Spain report (Izquierdo et al., 2013), which
present estimates of firm-level wage-growth equations with similar controls and show that wage residuals
appear to have been significantly smaller after the 2012 reform. Moreover, the available evidence suggests
that, while this wage moderation is affecting workers’ living standards, it is already yielding some
dividends in terms of employment dynamics. There is in fact a consensus among the few available
empirical studies published since the reform that, with respect to what is predicted by the negative GDP
32. Although formally the harmonized index of labour cost per hour worked is used as a dependent variable, to
the extent that controls capturing productivity developments are included in all specifications, the
estimated effect can be interpreted as estimates of the impact on unit labour costs (excluding non-wage
costs). Moreover, even though labour shedding is likely to be behind part of the effect visible in raw data
(cf. Figure 8), since the marginal worker who become unemployed is likely to be the one with the highest
wage with respect to his/her productivity, this is unlikely to be the case for the econometric estimates, to
the extent that cyclical fluctuations are rigorously controlled for in several ways in the regression models
(see Box 1 and the Annex).
**
**-2.0%
-1.8%
-1.6%
-1.4%
-1.2%
-1.0%
-0.8%
-0.6%
-0.4%
-0.2%
0.0%
Minimum Maximum
![Page 30: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/30.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 29
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
growth observed in the past six quarters, net employment contraction was below expectations, suggesting
therefore a positive impact of the reform on employment growth. Consistent findings on this issue emerge
from the simple comparison of the time series of GDP and employment growth (Ministerio de Empleo y
Seguridad Social, 2013), the examination of the residuals of an estimated Okun’s law (Izquierdo et al.,
2013) and the estimation of structural models (De Cea and Dolado, 2013, BBVA, 2013a). The further
improvement of employment trends in the third quarter of 2013 (see e.g. Puente and Font, 2013) seems to
confirm this consensus.
3.2. Changes in worker flows
Hiring rates
As discussed in Section 1.3, flexibility-enhancing reforms of hiring and firing regulations are, first
and foremost, expected to increase hiring and separations. However, it can be argued that the expected
impact of the 2012 reform on dismissals and separations is a priori ambiguous, due to the new incentives
to adopting internal-flexibility measures, thereby preserving jobs. Therefore, hiring behaviour by firms is
perhaps the most important area where the reform of hiring and firing regulations can be assessed.
Figure 10 underlines the dramatic contraction of the hiring rate during the crisis. Quarterly data from
the Labour Force Survey (Encuesta de Población Activa – EPA) show that the share of employees with
less than 3 months of tenure in total business-sector employment decreased substantially in 2008 (from
10.7% in the first quarter of 2008 to 7.4% in the first quarter of 2009, once corrected for seasonality).33 It
then recovered a little (up to 8.5%) in 2010 when the Spanish economy exhibited some timid signs of
recovery and then dropped again to 7.4% in the last quarter of 2011, pushed down by the worsening of
economic activity. Since then, the hiring rate has stabilised at 7%-8%.
To what extent can the stabilisation of the hiring rate in spite of the adverse cyclical conditions be
attributed to the 2012 reform? The estimated results obtained from hiring-rate regressions – which
carefully control for business-cycle influences and composition effects – cautiously suggest34 that the
reform had a moderate but positive impact on hiring (See Box 1 for the general methodology and the
Annex for details concerning all estimates presented in this section). Indeed, all other things equal the
reform is estimated to have increased the hiring rate by about 8% (that is raising the share of employees
with less than 3 months of tenure by about 0.6 percentage points),35 so that the hiring rate would have kept
falling until the beginning of 2013 in the absence of the reform (see Figure 10, Panel A).
The evolution of total hiring is mirrored by the dynamics of hiring on regular, open-ended contracts,
even though the latter typically represents only about one tenth of total hires. Again, the fall of hiring on
permanent contracts appears to have stabilised in 2012, followed by strong signs of an upward movement
in the second quarter of 2013.The econometric estimates suggest that the fall in permanent hiring would
33. Following OECD (2009, 2010), hiring rates are defined here as the share of workers hired in a given period
of time and that are in employment at the end of the period (counting therefore each worker only once)
divided by total wage and salary employment in that period. This definition is not uncommon in the
literature (see e.g. Davis et al., 2006, and Golan et al., 2006) and it is the most appropriate in the case of
data from the EPA.
34. It must be underlined however that, in contrast with other econometric analyses presented in this section,
the statistical significance of the estimated effect of the reform on the hiring rate depends on the
assumptions made on the polynomial trend used to capture business-cycle fluctuations (see the Annex).
35. More precisely, the microeconometric estimates show that the probability that an employee has job tenure
shorter than 3 months increased by about 0.6 percentage points. This probability is the micro-level
equivalent of the hiring rate at the aggregate level.
![Page 31: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/31.jpg)
30 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
have been more pronounced in the absence of the reform than what is predicted in the case of total hiring
(Figure 10, Panel B). In fact, the reform is estimated to have increased the hiring rate on permanent
contracts by 13%. In addition, the increase in permanent hiring in the aftermath of the reform appears to be
concentrated in full-time positions. In particular, the reform is estimated to have increased the hiring rate of
full-time open-ended contracts by 18% on average, while no significant effect is found as regards part-time
permanent contracts. By contrast, with an estimated positive effect of only 7% – insignificant in certain
specifications – the effect on hiring of temporary employees appears more limited.
Figure 10. The evolution of hiring rates in the non-agricultural business-sector, 2008-2013
Percentage rates corrected for seasonality
Panel B: Hires with a permanent contract
Note: The figure presents observed and predicted percentage ratios of workers with no more than 3 months of job tenure on total employment in the reference week. “Predicted without the reform” indicates the empirical prediction of what the hiring rate would have been in the absence of the reform, based on empirical estimates of the baseline model for the individual probability of having tenure equal to 3 months or less. Panel A refers to all new hires and panel B to new hires with a permanent contract.
Source: Estimations on the basis of data from the Encuesta de Población Activa (EPA). See Annex for the detailed estimation method and results.
0
2
4
6
8
10
12
Observed Predicted without the reform
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
Observed Predicted without the reform
![Page 32: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/32.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 31
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Transitions from unemployment to employment
Not only are flexibility-enhancing reforms of EPL expected to increase hiring, but they are also
expected to reduce unemployment duration and speed up transitions towards permanent employment.
Social security microdata, in which employment histories of individuals affiliated with social security can
be followed over time, are ideally suited for this analysis. For a large random sample of workers affiliated
with social security, the Muestra Contínua de Vidas Laborales (MCVL) registers all employment and
unemployment spells as well as the start and end dates, thereby measuring accurately their exact duration
and allowing to link these durations with previous employment spells. Figure 11 shows the proportion of
individuals leaving unemployment for permanent employment for a sample of Spanish workers aged 16-64
over the period 2006 to 2012, conditional on being unemployed at that point in time.36 The proportion of
unemployed individuals moving to a permanent job has decreased significantly over the period, in
particular due to the increase in the number of unemployed since the start of the crisis. However, there are
some signs of recovery in the most recent quarters. But to what extent can these developments be attributed
to the 2012 reform?
Figure 11: Average proportion of unemployed individuals leaving unemployment to permanent employment
Quarterly average of monthly rates, 2006-2012
Note: For each quarter, the figure presents the average monthly the proportion of individuals leaving unemployment to permanent employment for a sample of Spanish workers aged 16-64 over the period 2006 to 2012, conditional on being unemployed at that point in time. The sample includes all individuals who have been unemployed at least one day.
Source: Muestra Continua de Vidas Laborales (MCVL).
The estimation results indicate that the 2012 reform increased the probability of leaving
unemployment and entering employment for any unemployment duration.37 The effect of the reform is
36. The rates presented in Figure 11 are not comparable with those presented in Figure 10 for three main
reasons: i) averages of monthly (instead of quarterly) transitions are considered; ii) the denominator is
represented by the pool of unemployed workers (instead of wage and salary employees); and iii)
individuals not registered as unemployed are not included, which leaves out youth searching for their first
job and people transiting directly from job to job without spending at least one day in unemployment
(however the latter are less than 3% of all employees in the data).
37. Estimations of competing-risk hazard models on the MCVL – whose results are presented in this section –
were run by José Ignacio Garcia-Perez (University Pablo de Olavide) in close cooperation with the OECD
0.02
0.025
0.03
0.035
0.04
0.045
0.05
0.055
0.06
0.065
0.07
![Page 33: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/33.jpg)
32 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
significant only as regards the probability for the unemployed of being hired on a permanent contract,
which has increased by 24%, on average, during the first 6 months in unemployment where most
transitions occur (Figure 12). The estimated percentage effect is even greater at longer unemployment
durations (more than 40%), although obviously smaller in percentage points. By contrast, the effect of the
reform on transitions to a temporary contract is small and insignificant.38 Thus, for unemployment
durations no longer than 6 months, the reform appears to have raised the share of exits from unemployment
to permanent employment in total exits from unemployment by about 14%, taking the point estimates at
face value.
Figure 12: Exit from unemployment by type of contract and unemployment duration
Monthly probability of individual transition from unemployment to employment
Note: The figure presents the average of estimated monthly probabilities of individual transition from unemployment to employment just before (February 2012) and after (March 2012) the reform, by type of new contract and unemployment duration (less than six months, from 7 to 12 months and more than 12 months). The sample includes all individuals who have been unemployed at least one day. Unemployment durations are censored at 30 months in unemployment. Panel A refers to exits from unemployment to temporary contracts and Panel B to exits from unemployment to permanent contracts. The asterisks refer to the significance level of the estimated effect of the reform on each transition probability. **: significant at the 5% level.
Source: Estimations on the basis of data from the Muestra Continua de Vidas Laborales (MCVL). See Annex for the detailed estimation method and results.
Perhaps more important, the reform appears to have boosted transitions towards permanent
employment significantly more for those that had a temporary contract prior to becoming unemployed,
Secretariat, which however bears full and sole responsibility for the interpretation of the findings and the
choice of the modelling strategy.
38. However, in absolute terms, the estimated effect on transitions to permanent contracts is not significantly
greater than that on transitions to temporary contracts. For example, taking point estimates at face value, in
the first 6 months of unemployment, the average probability of a monthly transition to a temporary job
increased from 18.6% to 19.9% due to the reform whereas the average transition to permanent employment
increased only from 1.7% to 2.1%. Moreover, it cannot be excluded that these relative patterns are also due
to the gradual increase in severance pay for fixed-term contracts, starting in January 2012 but approved by
the parliament in 2010 (see Section 2.2).
0
0.05
0.1
0.15
0.2
0.25
6 months or less From 7 to 12 months More than 12 months
Before Reform After Reform
0
0.005
0.01
0.015
0.02
0.025
6 months or less From 7 to 12 months More than 12 months
A. Exit to temporary contract B. Exit to permanent contract
**
**
**
![Page 34: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/34.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 33
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
while the effect remains insignificant for those who previously held a permanent contract. Symmetrically,
the results suggest a marginally positive impact of the reform on the speed of transition towards a
temporary contract only for those that had a permanent contract prior to becoming unemployed. This is
particularly good news insofar as it yields evidence of greater mobility across contract types, suggesting
that the economy is on a slow path of a reduction of segmentation.
Separation rates
The overall effect of the reform on worker separations is less obvious, as the reform eased firing
procedures and reduced its costs but at the same time raised the incentives to adopt internal-flexibility
measures, in the first place to avoid terminations. The Encuesta de Coyuntura Laboral (ECL) is a large
establishment-level survey that registers worker flows at the establishment level and can, therefore, be used
to shed light on which effect dominates.
The estimation results show that the 2012 reform significantly reduced the proportion of workers
leaving the establishment in a given quarter. The average overall separation rate was reduced by 24% in
the aftermath of the reform (Figure 13). The reduction has been particularly sizeable for the separation rate
of part-time workers (41%) but there is also some evidence of an effect for full-time workers (13%), which
however is only marginally significant from a statistical point of view.
Figure 13: The effect of the reform on quarterly separation rates at the establishment level
Note: The figure shows empirical estimates of average separation rates in the post-reform period obtained from the estimation on quarterly data for the period 2006-2012 of the baseline model of establishment-level separation rates, as predicted by observable variables. “Predicted without the reform” indicates the empirical predictions of what separation rates would have been in the absence of the reform. For each establishment, separation rates are defined as the ratio of separations in a quarter divided by the average of total employment between the start and the end of the period. The asterisks refer to the significance level of the estimated effect of the reform on each separation rate. ***, **, *: significant at the 1%, 5% and 10% level, respectively.
Source: Estimations on the basis of data from the Encuesta de Coyuntura Laboral (ECL). See Annex for the detailed estimation method and results.
In addition, establishment-level data allows disaggregating flows by contract type and reason of
separation. In particular, it is possible to obtain distinct estimates of the effect of the reform on contract
terminations (for temporary workers), dismissals of permanent workers, terminations due to end of season
for seasonal workers with an open-ended contract – “fijos discontínuos” – and other reasons for separation
(including retirements, voluntary separations, deaths, etc.). The estimation results (Figure 14) point to a
***
*
**
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
Overall Full-time Part-time
Predicted without reform Predicted with reform
![Page 35: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/35.jpg)
34 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
significant impact of the reform on separations rates due to the end of contract for temporary workers (31%
decrease) and seasonal workers (71% decrease), but show no impact on dismissal rates and separations for
other reasons.
Figure 14. The effect of the reform on separation rates at the establishment level by type
Note: The figure shows empirical estimates of average separation rates in the post-reform period obtained from the estimation on quarterly data for the period 2006-2012 of the baseline model of establishment-level separation rates, as predicted by observable variables. “Predicted without the reform” indicates the empirical predictions of what separation rates would have been in the absence of the reform. For each establishment, separation rates are defined as the ratio of separations in a quarter divided by the average of total employment between the start and the end of the period. The asterisks refer to the significance level of the estimated effect of the reform on each separation rate. **: significant at the 5% level.
Source: Estimations on the basis of data from the Encuesta de Coyuntura Laboral (ECL). See Annex for the detailed estimation method and results.
The analysis of job spells and reasons for termination using social security microdata (MCVL)
appears to confirm the results for both dismissals and separations for other reasons obtained with
establishment-level data. For permanent workers, the estimation results do not show any effect of the
reform on the duration of job spells and the probability of separation, no matter whether for dismissal or
for other reasons. However, as regards fixed-term contracts, no effect of the reform on contract duration
and separation probabilities is found in MCVL data, in contrast with results obtained with establishment-
level data. Even though simple statistical reasons might explain this discrepancy,39 it suggests some caution
in interpreting these findings. At the very least, however, these results seem to confirm that the reform has
not increased the number of separations, in general, and dismissals, in particular. This fact suggests that the
stimulus provided by the reform to internal flexibility measures as an alternative to job suppression more
than compensated for the upwards effect on separations brought about by the easing of dismissal
procedures and costs.
39. For example, there is some evidence that the reduction of termination rates for temporary contracts is
concentrated in small establishments (see the Annex). As a consequence, the reform has reduced the
average establishment-level separation rate with smaller or no impact on aggregate separations (insofar as
aggregate separation rates can be seen as a weighted average of establishment-level ones, with
establishment size as weight).
**
**
0
0.01
0.02
0.03
0.04
0.05
0.06
End of Contract Dismissal End of Season Other
Predicted without reform Predicted with reform
![Page 36: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/36.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 35
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Individual and collective dismissals
Has the distribution of dismissals – between individual and collective terminations – been affected by
the reform? It is possible to resort to monthly statistics on inflows onto unemployment benefits
(prestaciones por desempleo) to investigate the relative variation of individual and collective dismissals,
insofar as the reason for claiming benefits is reported in these data.40 The share of collective dismissals
increased steadily since 2008, from about 5% to a peak of 20% in May 2013 (Figure 15). However, there is
no sign that the reform is responsible for this trend: the share of inflows onto unemployment benefits due
to collective dismissals seems to have been on the rise even before, probably as the result of the
restructuring of the economy brought about by the economic crisis. On the contrary, the estimated results
indicate that the share of collective dismissals would have been 3.1 percentage points larger in the absence
of the reform. In practice, the empirical estimates suggest that the 2012 reform decreased the number of
inflows onto unemployment benefits for reasons of collective dismissals by about 32%, while in the same
period the contraction of inflows due to individual dismissals was small and marginally significant.41
Figure 15. Share of inflows onto unemployment benefits due to collective dismissal in total inflows due to dismissal, 2007-2013
Monthly percentage share of inflows
Note: The figure presents the monthly percentage share of inflows in unemployment benefits due to collective dismissal in total inflows due to dismissal, as observed in raw data and as they would have been in the absence of the reform, according to baseline estimates on the basis of data covering the period January 2007-July 2013.
Source: Estimations on the basis of data on the prestaciones por desempleo de nivel contributivo (Source: MEYSS). See Annex for the detailed estimation method and results.
40. However, some cautions must be exerted in interpreting the results from this analysis since not all
dismissals are counted. Inflows data considered here are limited to workers who are eligible to standard
unemployment benefits because of sufficient previous contributions and contribution periods (Prestaciones
de nivel contributivo). They exclude, therefore, workers only eligible for other types of unemployment-
related subsidies (Prestaciones de nivel asistencial). Moreover, inflows data considered here exclude those
resulting from conciliation settlements, court rulings, end of trial period or contract suspensions, because
the nature of the originating event (individual or collective dismissal) is not reported. By contrast, benefit
claims ex Ley 45/2002 – individual dismissals for which benefits are claimed before any court ruling
stating whether the dismissal is fair or unfair – are included in the analysis.
41. In the baseline model, the estimated contraction for these inflows is of 6%. However this estimate is
significant at the 10% level only and is not robust to changes in the specification (see the Annex for more
details).
0
5
10
15
20
25
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
2007 2008 2009 2010 2011 2012 2013
Observed Predicted withouth the reform
![Page 37: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/37.jpg)
36 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Taking into account the discussion of Section 2.2, these findings are probably not surprising. There is
little controversy that the 2012 reform reduced the cost and difficulty of individual dismissals, which
explains the pick-up on hiring on permanent contracts that has been observed since its implementation (see
above). As already noted, the lack of impact of the reform on individual dismissals is likely to be due to the
counterbalancing effects of greater use of internal flexibility measures and lower cost of individual
dismissals. By contrast, the case of collective dismissals is more complex. On the one hand, the reform has
made economic dismissals significantly easier, even though this is true for both individual and collective
dismissals. On the other hand, the reform, under certain circumstances, has increased the taxes levied on
firms in the case of collective dismissals (see section 1). Moreover, the reform, as implemented initially,
significantly increased the risk that the dismissal procedure could be invalidated by a court in the case of
litigation. The observed empirical patterns tentatively suggest that these factors are likely to have played a
key role in discouraging collective dismissal decisions.42 Yet, to the extent that the Real Decreto Ley
11/2013 approved in August 2013 has restricted the discretionary role of courts of declaring a dismissal
void and ordering reinstatement (see Section 1), it would not be surprising if some of the negative effects
of the 2012 reform on the share of collective dismissals in total dismissals are subsequently unwound.
Firm-size differences in the impact of the 2012 reform
Given the importance of the provisions targeted at small firms in the architecture of the reform (see
Section 1), it is important to evaluate firm-size differences in its impact. In Spain, employers have the
obligation to inform the public employment service about the characteristics of all new individual contracts
they sign as well as about extensions of fixed-term contracts and/or their conversions into permanent
contracts. These data are then collected into a database – containing also workers and employers
characteristics, including firm size – that can be used to analyse the dynamics of new contracts on a
monthly basis, allowing the examination of firm-size heterogeneity in hiring patterns.
In January 2007, about 250,000 new permanent contracts were signed (Figure 16). Of these, about
two-thirds were in firms with 50 or fewer employees (cf. Panel A and B). As already noted, the crisis hit
job creation hard. Five years later, new permanent contracts were as few as 75,000 per month, with the
proportion accounted by small firms remaining approximately constant. The trends concerning new open-
ended contracts by firm size has diverged since then. For example, in January 2013, firms with 50
employees or less accounted for more than three-fourths of all new permanent contracts concluded that
month.
To what extent are these developments accounted for by the 2012 reform? The econometric models
estimated for this report suggest that at least 25 000 new permanent contracts per month in firms with 50
employees or less can be attributed to the 2012 reform, while no significant effect is observable for larger
firms. These patterns are mirrored by those of the share of permanent contracts in all new contracts, which
is estimated to have been raised by 3 percentage points – that is a percentage increase of about 30% – for
firms with 50 or fewer employees.
42. However, as discussed in Section 1, individual dismissals have become particularly easier and less costly in
the case of small firms: companies with less than 50 employees can hire by stipulating an open-ended
contrato emprendedores, which allows a trial period of up to one year, while companies with less than
25 employees can apply for a severance-pay subsidy for any type of fair dismissal. It is not impossible that
part of the effect of the reform on the share on collective dismissals can be attributed to these provisions
targeted at small firms. As benefit inflows disaggregated by firm size are not available, this issue cannot be
investigated further here. Firm size heterogeneity in the impact of the reform is examined in the next
subsection as regards other types of flows.
![Page 38: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/38.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 37
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Figure 16. Monthly inflows into permanent contracts, including contract conversions, 2006-2013
Panel A: Firms with 50 or fewer employees
Panel B: Firms with more than 50 employees
Note: The figure presents the number of new permanent contracts signed each month, as observed in raw data and as they would have been in the absence of the reform, according to baseline estimates covering the period January 2006-July 2013.
Source: Estimations on the basis of data on the datos estadísticos de contratos (Source: SEPE). See Annex for the detailed estimation method and results.
Consistently, it appears from the analysis of social security microdata that the 2012 reform had a
greater impact on transitions from unemployment to permanent employment in small establishments rather
than large establishments (Figure 17).43 The results from the estimation of the competing-risk hazard
43. Although a disaggregation by firm rather than establishment size would be more interesting, this is not
possible given the available data. The only size variable which is available in the Muestra Continua de
Vidas Laborales is the size of the social security accounting unit (Cuenta de Cotización), which is unique
for each firm at the provincial level. It does not correspond exactly either to the plant or the firm level – it
covers many plants of the same firm if they are in the same province but plants established in different
0
20000
40000
60000
80000
100000
120000
140000
160000
180000
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
2006 2007 2008 2009 2010 2011 2012 2013
Observed Predicted withouth the reform
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
100000
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
2006 2007 2008 2009 2010 2011 2012 2013
Observed Predicted withouth the reform
![Page 39: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/39.jpg)
38 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
models suggest that the reform increased the average transition to a permanent contract – conditional on
being unemployment for six months or less – by 26% in establishments with 50 or fewer employees,
compared with 15% in establishments with over 50 employees. To the extent that the large majority of
firms are mono-establishments, this confirms that the impact of the reform on hiring is greater for smaller
firms.
Figure 17. Exit from unemployment to permanent employment by firm size
Monthly probability of individual transition for unemployment durations smaller or equal than 6 months
Note: The figure presents the average of estimated monthly probabilities of individual transition from unemployment to permanent employment just before (February 2012) and after (March 2012) the reform conditional on being unemployed at least one day but less or equal than 6 months, by type of contract and firm size. Unemployment durations are censored at 30 months in unemployment. The asterisks refer to the significance level of the estimated effect of the reform on each transition probability. ***: significant at the 1% level.
Source: Estimations on the basis of data from the Muestra Continua de Vidas Laborales (MCVL). See Annex for the detailed estimation method and results.
When the data are further disaggregated by firm size, the greatest estimated expansion in the number
of new permanent contracts induced by the reform is found in very small firms with 25 employees or less
(45%, corresponding to about 21,000 new contracts per month). By contrast, in firms with 26 to
50 employees the increase was relatively more modest, not only in absolute terms (about 4,000 new
contracts per month) but also in percentage terms (about 28%).
Given that the positive effect of the reform appears concentrated in very small firms, a first obvious
candidate in explaining these findings is the severance-pay subsidy made available for firms with less than
25 employees in all cases of fair dismissal (see Section 1). Indeed, to the extent that fair economic
dismissals have become easier and less uncertain after the reform, this subsidy is likely to play a greater
role than before the 2012 reform, when most employers used to take the route of “despido exprés” and
severance payments were therefore not subsidised. Obviously, however, this provision alone cannot
explain the significant impact for larger companies. Another factor that is likely to be behind this upsurge
provinces are not linked. For a matter of simplicity each social security accounting unit is referred to as an
“establishment”, hereafter.
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
Size<=50 Size>50
Before Reform After Reform
***
***
![Page 40: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/40.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 39
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
in hiring on permanent contract is the introduction of the contrato emprendedores, which de facto extended
the trial period for firms with less than 50 employees.44 Indeed, for these firms, the positive effect of the
reform is somewhat larger on new hires than on contract conversions, which in the period of analysis
concerned, by and large, correspond to workers hired before the reform. Yet, some significant impact of
the reform is detectable in the case of conversions as well.
Can these two factors, taken together, fully explain the firm-size heterogeneity of the effect of the
reform? This is far from clear. Less than 8,000 contrato emprendedores are typically stipulated in a
representative month. Given that the number of new permanent contracts in firms with 25 or fewer
employees is on average 5 times larger than the corresponding figure for employers with 26 to
50 employees, a large share of the new contrato emprendedores is likely to have been signed in firms with
25 or fewer employees.45 Moreover, in firms with 26 to 50 employees, the impact of the reform on contract
conversions is very close to that on new hires on permanent contract. These observations suggest a third
factor that could be behind the observed empirical patterns. The larger the size of the firm, the greater the
likelihood the firm needs to undertake a large number of dismissals at some point in time. But the threshold
triggering collective dismissal procedures varies little with firm size (from 10 dismissals within a period of
3 months for firms with less than 100 employees to 30 employees within the same period for firms with
300 employees of more); therefore, collective dismissals procedures are likely to be more often binding in
the case of larger firms.46 Given that the evidence presented in the previous subsection and in Section 2.2
suggests that collective dismissals might not have become less difficult and costly, larger firms might have
refrained from hiring on permanent contracts to avoid incurring the risk of making collective dismissals in
the future, particularly at a time of sluggish demand and uncertain perspectives.
This conjecture can be tested by examining further by more disaggregated size classes the impact of
the 2012 reform on hiring by firms with more than 50 employees. The estimated results based on contract
data show that the reform has significantly increased the number of new permanent contracts in firms with
51 to 100 employees (by 23%), while no significant effect is detected for larger firms.47 To the extent that
all these firms are ineligible both for the contrato emprendedores and for the severance-pay subsidy, the
combination of easier individual dismissals and more burdensome collective dismissals remains the only
explanation for the simultaneous increase in the number of permanent contracts stipulated by firms with
51 to 100 employees and the lack of significant changes for larger firms. Moreover, for the former class of
firms, the positive effect of the reform is entirely concentrated in contract conversions, suggesting that
these firms use fixed-term contracts as a substitute for longer trial periods.
Overall the evidence presented in this section not only suggests that the Spanish labour market has
become more dynamic in the aftermath of the 2012 reform but also that duality is decreasing. Indeed,
hiring on permanent contracts is on the rise while there is no evidence of an increase of transitions out of
permanent contracts. The latter fact is also confirmed by recent evidence provided by the Ministerio de
Empleo y Seguridad Social (2013), which shows that there is no difference in the probability of leaving a
44. This could also contribute to explain why the effect is concentrated in very small firms: to the extent that
mismatching problems are likely to be more problematic in these firms, they can benefit more from the
longer trial period made available by the contrato emprendedores.
45. No breakdown by firm size is available for the contrato emprendedores.
46. For example, according to OECD (2009) the average quarterly dismissal rate of continental European
countries for which data are available is close to 1%. That is, in a normal year, almost 1% of employees are
dismissed each quarter. Therefore, any firm with 3,000 employees or more dismissing at that rate would
cross the threshold defining collective dismissals.
47. The difference between these two classes of firms in the effects of the reform is also significant, at least at
the 10% level.
![Page 41: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/41.jpg)
40 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
permanent contract conditional on tenure between a contrato emprendedores and a standard open-ended
contract in the first 12 months of a job spell. This implies that longer trial periods in the former type of
contract have not implied greater precariousness for workers hired with these contracts. Moreover,
dynamic segmentation has also been reduced insofar as transitions between temporary and permanent
contracts have become more frequent (see above). Nevertheless, if the percentage increase of transition
rates that can be attributed to the reform is sizeable, their absolute magnitude is small. Moreover, there is
some evidence suggesting that the reform have particularly reduced separations for temporary contracts,
due to the greater incentives for internal flexibility as an alternative to the suppression of fixed-term posts
(see Figure 14). As a consequence, it is likely that it will take some time before these developments
become visible in the share of temporary employees in total employment. Indeed, using the same
microeconometric model used before for hiring rates but replacing the dependent variable, no significant
effect on the share of workers with a fixed-term contract is estimated if the regression discontinuity is set at
the time of the reform (first quarter of 2012),48 while a modest – albeit robust and significant – effect
appears if the discontinuity is tested one year later.49 It is, however, too early to say what will be the
magnitude of the long-run impact of the reform on this variable.
3.3. Simulating the long-run impact of the 2012 reform on productivity and economic growth
The increase in the pace of reallocation of labour resources and the reduction in labour market duality
are expected to increase efficiency and labour productivity growth in the long term (see Section 2.2).
However, insofar as multi-factor productivity growth – measuring efficiency enhancement in an economy
– cannot be meaningfully estimated at greater frequencies than annual ones, it is not possible to test
whether the 2012 reform has effectively brought about this outcome. Nevertheless, by exploiting the
quantification of the institutional change implied by the reform – as reflected in OECD EPL indicators –
and resorting to estimates available in the literature, it is possible to provide a tentative estimate of the
potential impact of the 2012 reform on productivity growth and, with some additional assumptions, overall
economic growth.
Bassanini et al. (2009) estimate the long-run potential impact of changes in the OECD summary
indicator of EPL for regular contracts (see Section 2.2) on multi-factor and labour productivity growth in
the business sector, excluding agriculture, mining, fuel and professional services.50 Their industry-specific
estimates are based on the characteristics of each industry, notably as regards the propensity to use
dismissals as a way to adjust to shocks.51 Taking their most reliable estimate at face value and assuming as
given the sectoral structure of the Spanish economy in 2007,52 in the long-run an hypothetical reform
reducing the EPL indicator for individual and collective dismissals by half a point – a large reform in terms
of the historical record for OECD countries –53 would result, in the business sector, in higher multi-factor
48. This finding is confirmed by running the same estimation model on social security data.
49. The estimated drop of the share of temporary workers is of 0.5 percentage points.
50. These industries were excluded due to the difficulty of measuring accurately multi-factor productivity
growth therein.
51. This propensity is approximated by industry-specific US dismissal rates, a choice justified by the light
firing regulations in this country that can be then taken as a benchmark to mimic the distribution of
dismissal rates in the absence of regulation.
52. This year is chosen to avoid that the simulation be perturbed by large cyclical swings in the structure of the
economy. Results are however robust to the choice of the year.
53. The 2003 severance-pay reform in Austria, often cited as an example of large, radical reform, resulted in
about half a point reduction of these indicators.
![Page 42: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/42.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 41
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
annual productivity growth by 0.45 percentage points and faster labour productivity growth by
0.3 percentage points.
What predictions would these estimates imply for the Spanish economy? The size of the 2012 reform,
as measured by OECD indicators, is only slightly smaller (0.44; cf. Figure 4) than what considered by
Bassanini et al. (2009). By applying the same estimated coefficients, it can be expected that, in the long-un,
labour productivity should grow faster by about ¼ of a percentage point each year in the business sector
(excluding agriculture, mining, fuel and professional services) as the result of the reform. Taking into
account that these industries accounted for 59% of total value added in the Spanish economy before the
crisis, and assuming conservatively no impact on employment and productivity in other industries, this
would translate into 0.15 faster GDP growth each year. This number can be seen as very rough lower-
bound estimate, to the extent that it is based on the assumption of no impact on other industries and no
impact on employment. While the latter assumption would be reasonable had the reform been limited to
employment protection, the wage moderation induced by the reform of collective bargaining is likely to
result in greater employment growth, thereby raising GDP growth even further.
4. Interactions with other policies
The previous section has assessed the impact of the 2012 reform of Spain’s labour market legislation
as has occurred so far. However, these reforms should be considered and evaluated in the broader context
of structural reforms and policies in other areas. Indeed, as suggested by the Revised OECD Jobs Strategy
(OECD, 2006), the effects of macroeconomic and structural policies are strongly interdependent and
institutions should be designed and reformed in such a way to exploit their complementarity. In other
words, the full materialisation of the positive employment impact of the recent labour market reform is
strictly conditional on complementary reforms in other areas.
First of all, macroeconomic conditions affect overall labour market performance and the effect of the
2012 reform will depend on any macroeconomic shocks the Spanish economy might suffer. Sound
macroeconomic policies are necessary to support economic and employment growth. In that respect, it is
important to keep the public debt on a sustainable path, while carefully avoiding excessively restrictive
fiscal stances. The pension reform proposed by the government, following the recommendations of a
commission of experts, would go in this direction, as it would ensure the long-term sustainability of the
pension system as well as intergenerational solidarity. Indeed the reform proposal would envisage
coupling, by 2019, the introduction of a revalorisation index (Índice de Revalorización de las Pensiones) –
guaranteeing the adjustment of pensions only conditionally on the favourable evolution of a large number
of parameters – with that of a sustainability factor (Factor de Sostenibilidad) – linking the amount of the
retirement pension to changes in life expectancy (see e.g. BBVA, 2013a).
In addition, numerous interactions exist between labour legislation and product market institutions.
Reforms in the product and service markets will not only increase competition and boost productivity, but
they will also have an impact on labour market performance (Blanchard and Giavazzi, 2003, Nicoletti and
Scarpetta, 2005). Spain has already undertaken several reforms in 2012 and 2013 (e.g. Programa Nacional
de Reformas 2013 and the update of the Stability Program Update for Spain 2013-2016). However, it is not
yet clear to what extent the fall in unit labour costs (see Section 3.1) translates into a commensurable fall of
price inflation54, thereby guaranteeing a full impact of wage moderation on competitiveness. To avoid that
slower wage growth results only in higher profits, without raising competitiveness, a further push on
54. For example, the average increase in wages as set in collective agreements in 2012 was 1.3% (0.9% in new
collective agreements), while inflation was 2.9% (Izquierdo et al., 2013), a relative high level in
comparison with Eurozone competitors. However figures for 2013 suggest that inflation is declining
significantly and is now close to 0.
![Page 43: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/43.jpg)
42 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
product and service market competition might be necessary (European Commission, 2013). Estimates by
Bassanini and Duval (2009) suggest that the effect on unemployment of simultaneous reforms of anti-
competitive product market regulations and industrial-relation institutions – such as the reform of
collective bargaining discussed in Section 3.1 – is at least 15% larger than the sum of the effects of the two
policy reforms taken in isolation.
The reform of the financial sector is currently underway and remains on track (IMF, 2013).
Nevertheless, debt levels remain high and constrain the private sector’s ability to borrow (La Caixa, 2013).
Available evidence from other countries suggests that falling bank lending dramatically affects investment
of small and medium enterprises, making their employment level slump. For example, Chodorow-Reich
(2014) finds that the withdrawal of credit accounted for between one-third and one-half of the employment
decline in US small and medium firms in the year following the Lehman bankruptcy. Similarly, using
Italian data, Cingano et al. (2013) find that a 10 percentage-point fall in credit growth reduces the
investment rate by 8-14 percentage points over four years, and employment by almost 1.5%. Restoring the
health of the financial markets and easing the credit crunch is particularly important in the case of Spain,
given the large share of small firms in this economy.
Another essential element to improve the functioning of the Spanish labour market and reviving
employment creation relate to active labour market policies. In Spain, the administration of unemployment
benefit system depends on the Spanish public employment service (SEPE) in the central government, while
the development of active labour market policies is done by the regional governments (following the
legislative framework approved by the central government).
The Spanish unemployment insurance is relatively generous with respect to the OECD average
(Figure 18). Providing adequate unemployment benefits can be desirable in the context of significant
structural reforms: while many workers are likely to gain from these reforms, certain groups of workers
would inevitably lose their jobs in their aftermath and require support to regain employment. However
unemployment benefits must be made conditional on strictly-enforced work-availability conditions and
included in a well-designed “activation” package. The decentralized public employment service may not
have the capacity for implementing activation policies on a sufficient scale to counterbalance the potential
negative effects of generous benefits on work incentives. Spain’s expenditure on passive measures is
around 3% of GDP (the highest level in the OECD), while expenditure on active measures is slightly
below 1%,55 and most of the “active” expenditures are employment incentives. More important, the
plurality of actors in this area makes it difficult to integrate properly-enforced job-search requirements and
effective re-employment services with the provision of adequate unemployment benefits. Last but not least,
resources for preventing the receipt of benefits when in undeclared employment have always been limited
in Spain.
Since the start of the recession, there has been a movement along the Beveridge curve (Figure 19),
which allows identifying changes in the efficiency of the process of matching unemployed workers to
vacancies – an inward shift of the curve representing an improvement in matching efficiency. The figure
shows a decrease in the vacancy rate paralleled by an increase in unemployment, corresponding to the
increase in labour market slack. Yet, no inward shift of the curve is detectable since the 2012 labour
market reform, suggesting that inefficient activation policies are hampering the improvement in matching
efficiency. In the most recent quarters, there seems to be even an outward shift denoting intensification in
matching frictions, although probably due only to the simultaneous increase in long-term unemployment
and the renewed firms’ propensity to post vacancies (see Section 3.2).
55. Even if part of the low ratio between active and passive expenditures is due to the recession.
![Page 44: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/44.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 43
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Figure 18. Average of net replacement rates over the first 60 months of unemployment, 2011
Unweighted average of two income levels and four family situations, excluding social assistance
Note: Unweighted averages, for full-time earnings levels of 67% and 100% of the average wage (AW) and four family situations (single persons, couple, with two children and without). Family benefits are included. Any income taxes payable on unemployment benefits are determined in relation to annualised benefit values (i.e. monthly values multiplied by 12) even if the maximum benefit duration is shorter than 12 months. For married couples the percentage of AW relates to the previous earnings of the "unemployed" spouse only; the second spouse is assumed to be "inactive" with no earnings and no recent employment history. Children are aged four and six and neither childcare benefits nor childcare costs are considered.
Source: OECD Tax-Benefits Models.
Figure 19. The Beveridge curve in Spain
Vacancy and unemployment rates as a percentage of the labour force, Q1 2001-Q2 2013
Note: Job vacancy rates refer to the non-agricultural sector. The light blue line corresponds to the period up to 2007 Q4, while the dark blue line corresponds to the period since 2007 Q4.
Source: OECD calculations based on OECD Short-Term Labour Market Statistics Database and Eurostat, Job Vacancy Statistics.
0
10
20
30
40
50
60
70
Q1 2001
Q2 2013
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
7 9 11 13 15 17 19 21 23 25 27
Vaca
ncy
rate
(%
)
Unemployment rate (%)
![Page 45: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/45.jpg)
44 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Some improvements have taken place in the domain of activation policies. The strategy for
entrepreneurship and youth employment 2013-2016 (Estrategia de Emprendimiento y el Empleo Joven
2013-2016) approved in February 201356 includes measures to encourage youth employment, facilitating
job-placement in salaried employment as well as the start of entrepreneurship activities. For example,
among the measures approved so far, new hiring incentives to employ unemployed youth have been
created and a reduced flat rate for social security contributions for young entrepreneurs has been
introduced. Moreover, the annual coordination mechanism between central and regional governments
(Plan annual de política de empleo 2013)57 approved in August 2013 has introduced several changes to
improve the coordination between national and regional administrations. This agreement states the annual
objectives of employment policies in Spain and introduces several evaluation measures of the activities
performed by the regional governments. Up to 40% of the funding of active labour market policies of
regional governments will be conditional on those evaluation measures. In addition, since August 2013, an
individual needs to be registered as unemployed in the regional office and keep this status in order to
receive and maintain the right to receive unemployment benefits.
The changes introduced in the domain of active labour market policies go in the right direction, but it
remains to be seen whether they are sufficient to be effective, in particular insofar as they fall short of fully
integrating active and passive policies. In the context of a prolonged recession, it is very important to focus
on improving activation policies to help those unemployed to find jobs. For this, it is crucial to have the
right institutional setting with effective employment services. Further in-depth analysis would be required
in order to identify the possible weaknesses in the current Spanish system in the delivery and provision of
income support and employment services. In particular, it will be important to ensure transparency in the
monitoring of the effectiveness of regional active labour market programmes and that there are strong
incentives at the local level to follow best practice. This is the case in Switzerland, for example, where
cantons work under very clear guidelines and are subject to close scrutiny by the central government (see
e.g. Düll et al., 2010, OECD, 2013a).
5. Conclusions and recommendations
The 2012 labour market reform in Spain appears to have been a significant step in the right direction.
As the result of it, the Spanish labour market has already shown some signs of increased dynamism and
this is likely to bring about lower dualism and faster productivity growth in the medium term. The reform
focused primarily on collective bargaining and on dismissal regulations. This report provides a very
preliminary assessment of the short-run effects of the reform on labour market performance. This is an
important caveat insofar as: i) only a short period has elapsed since the implementation of the reform; ii)
Spain is only now showing timid signs of recovery from a long recession; and iii) due to the breadth of the
reform, it is difficult to clearly identify a counterfactual to carry out a standard evaluation. As a
consequence, further monitoring is required to fully assess the impact of the reform in the medium and
long term.
The analysis presented here confirms that, as suggested elsewhere (e.g. Izquierdo et al. 2013,
Ministerio de Empleo y Seguridad Social, 2013, BBVA, 2013a), the effect of the reform on internal
flexibility and collective bargaining has played an important role in the recent contraction of unit labour
costs, although protracted adverse cyclical conditions are not alien to this pattern. It is not yet possible to
say whether the reform will restore competitiveness over the medium to long term. However, the Spanish
economy appears to have made progress in achieving wage moderation, which allows setting jobs in times
of crisis, even if continuous monitoring of the effect of the reform in this area is advisable, and the
56. Real Decreto-ley 4/2013, full legislative text available at
http://www.boe.es/boe/dias/2013/02/23/pdfs/BOE-A-2013-2030.pdf.
57. The full legislative text is available at http://www.boe.es/boe/dias/2013/09/10/pdfs/BOE-A-2013-9464.pdf.
![Page 46: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/46.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 45
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
government must be ready to implement further action if performance worsens. In addition, trends in
income inequality should be monitored in order to guarantee that the costs and benefits of the reform are
equally shared.
The analysis also shows an effect of the reform on hiring, particularly on permanent contracts, which
is most likely the outcome of the relaxation of dismissal regulations. Consistently, the reform appears to
have also reduced the duration of unemployment spells, essentially due to faster transitions towards a
permanent contract. In addition, there are signs of a fall in the segmentation of the labour market insofar as
transitions to permanent jobs have increased in particular for workers entering unemployment after a
temporary contract. There is also some evidence of a reduction in separations, particularly for temporary
workers, possibly resulting from the application of internal flexibility measures as an alternative to contract
termination. Nonetheless, there is also evidence of a negative effect of the reform on collective dismissals,
probably due to the increase in procedural uncertainty that initially followed the elimination of the
administrative authorisation as well as the increase in the relative direct costs of collective dismissals with
respect to individual terminations. In this respect, the government was quick to react with a mini-reform in
August 2013 to reduce this uncertainty. It is too early to say whether this intervention will prove sufficient
to restore outcome predictability. Yet, the discretionary role of courts to invalidate a collective-dismissal
procedure and order reinstatement of workers with backpay remains large, at least on paper.
The increase of new permanent jobs is concentrated in small and medium sized employers, despite the
persistent credit crunch, which is likely to affect especially this class of firms, insofar as they are more
reliant on bank lending than on equity finance or corporate bonds. The greater burden on collective
dismissals can be considered as one of the factors that explain why no increase in hiring is observed for
large firms. Nonetheless, two other elements appear to have played a key role in determining the faster
growth of open-ended contracts in small firms: i) the extension of the trial period for firms with less than
50 employees, made possible by the introduction of the contrato de Apoyo a emprendedores; and ii) the
severance-pay subsidy, made available to firms with less than 25 employees in all cases of fair dismissal.
The latter is likely to play a greater role after the 2012 reform, insofar as fair economic dismissals have
become easier.
The results of the empirical analysis suggest that the reform could have contributed to
25 000 additional new permanent contracts each month and increased the share of permanent jobs in new
hires by 3 percentage points. However, the share of temporary contracts in new contracts remains high.
Moreover, there is some evidence that incentives to internal flexibility measures have particularly
contributed to save the jobs of workers on fixed-term contracts. Thus, these pieces of evidence suggest that
it may take some time before these new patterns have a visible impact on the proportion of fixed-term
employees in total employment.
Whether or not the 2012 labour reform is sufficient to transform the Spanish labour market into one
that combines flexibility with fairness and worker security remains to be seen. Nevertheless, to improve the
likelihood that this objective will be attained, the empirical findings presented in this report suggest that
further adjustments might be desirable and could be put in place at little social cost.
First, given the evidence presented in this report, some action as regards regulation for collective
dismissals could be taken. Even though it cannot be established whether the August 2013
revision is sufficient to rebalance incentives, a cautious approach would suggest reducing further
the discretionary role of courts to invalidate dismissals, restricting it only to cases of
discrimination and prohibited grounds – as in the case of individual terminations. In this way,
unlawful collective dismissals will be treated – as in most other OECD countries – as any other
type of unfair dismissal, thereby restoring the equilibrium between individual and collective
dismissals. Moreover, some of the additional, specific costs for employers in the case of
![Page 47: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/47.jpg)
46 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
collective dismissals introduced by the reform could be rolled back, notably those contributions
linked to the firm’s future profitability, which increase uncertainty at the time of starting the
dismissal procedure about the effective financial burden to be borne by employers.
Second, there is evidence that the trial period is still too short for most firms that are not eligible
for the contrato emprendedores. Firms with 50 to 100 employees, not eligible for this type of
contract, have increased their share of permanent contracts since the 2012 reform. However, this
increase is entirely due to contract conversions, showing that these employers still tend to use
fixed-term contracts as a surrogate for longer trial periods in order to screen new recruits. To the
extent that dismissing non-performing or mismatched employees after the expiration of the trial
period remains difficult in Spain, the government could therefore consider lengthening the
maximum duration of trial periods up to at least the OECD average for firms not covered by the
contrato emprendedores, particularly in those cases in which this duration is currently much
shorter.
Third, a greater convergence of employers’ costs of termination for permanent and temporary
contracts would be desirable. This will be already the case by 2015 for firms with less than 25
employees – for fair terminations – due to the severance-pay subsidy in place for these firms and
the schedule of increases in severance pay for fixed-term contracts approved in 2010. By that
date, for firms within this size class, severance costs borne by the employer for fair terminations
at its own initiative will be equalised, no matter whether the contract is open-ended or fixed-term.
The government could consider deepening this convergence process by reducing severance pay,
and in particular ordinary severance costs for large employers.
Last but not least, greater integration of active and passive policies is required in order to increase
the capacity of the Spanish economy of matching the unemployed with vacancies. In this respect,
a careful evaluation of recent reforms in this area is required. In particular, it will be important to
ensure transparency in monitoring the effectiveness of regional active labour market programmes
and that there are strong incentives to follow best practice. In addition, the introduction of a
Youth Guarantee scheme as recommended by the European Council earlier this year for all
EU countries would also be a welcome step to reduce unemployment and NEET (not in
employment or education and training) rates among youth. But this will require ensuring that
cost-effective public and private employment services are in place to roll out the guarantee while
maintaining a high level of services for all job seekers.
![Page 48: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/48.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 47
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
REFERENCES
Babecký, J., Du Caju, P, Kosma, T., Lawless, M., Messina, J. and T. Rõõm (2009), "Downward Nominal
and Real Wage Rigidity: Survey Evidence from European Firms," ECB Working Paper Series 1105,
European Central Bank, Frankfurt.
Bassanini, A. and A. Garnero (2013), “Dismissal Protection and Worker Flows in OECD Countries:
Evidence from Cross-country/Cross-industry Data”, Labour Economics, Vol. 21, pp. 25-41.
Bassanini, A. and R. Duval (2009), “Unemployment, institutions and reform complementarities: Re-
assessing the aggregate evidence for OECD countries”, Oxford Review of Economic Policy, 25(1),
40-59.
Bassanini, A., L. Nunziata and D. Venn (2009), “Job Protection Legislation and Productivity Growth in
OECD Countries”, Economic Policy, Vol. 58, pp. 349-402.
BBVA (2013a), Spain Economic Outlook, Second Quarter 2013, BBVA Research, Madrid.
BBVA (2013b), Flash España, 28 de febrero de 2013, BBVA Research, Madrid.
BBVA (2013c), Spain Economic Outlook, Fourth Quarter 2013, BBVA Research, Madrid.
Bentolila S., P. Cahuc, J. Dolado and T. Le Barbanchon (2012), “Two-Tier Labour Market Reforms in the
Great Recession: France vs. Spain”, Economic Journal, vol. 122(562), pages F155-F187.
Bentolila S., J. Dolado and J. Jimeno (2008), “Two-Tier Employment Protection Reforms: The Spanish
Experience”, CESifo DICE Report, Vol. 6, pp. 49-56.
Blanchard, O. and F. Giavazzi (2003), "Macroeconomic Effects Of Regulation And Deregulation In Goods
And Labor Markets," The Quarterly Journal of Economics, MIT Press, vol. 118(3), pp. 879-907,
August.
Boeri, T. and P. Garibaldi (2007), “Two Tier Reforms of Employment Protection: A Honeymoon Effect?”,
Economic Journal, Vol. 117, pp. F357-F385.
Brändle, T. and W.D. Heinbach (2010), "Opening Clauses in Cellective Bargaining Agreements: More
Flexibility to Save Jobs?," IAW Discussion Papers 67, Institut für Angewandte
Wirtschaftsforschung (IAW).
Cappellari, L., Dell’Aringa, C. and M. Leonardi (2012), "Temporary Employment, Job Flows and
Productivity: A Tale of Two Reforms," Economic Journal, vol. 122(562), pages F188-F215.
Card, D. and S. Lee (2008), "Regression discontinuity inference with specification error," Journal of
Econometrics, Elsevier, vol. 142(2), pages 655-674, February.
Chodorow-Reich, G. (2014), The Employment Effects of Credit Market Disruptions: Firm-level Evidence
from the 2008-09 Financial Crisis, The Quarterly Journal of Economics 129, forthcoming.
![Page 49: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/49.jpg)
48 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Cingano, F., F. Manaresi and E. Sette (2013), “Does Credit Crunch Investment Down?”, Bank of Italy,
mimeo.
Dahl, C.M., D. le Maire and J. R. Munch (2011), “Wage Dispersion and Decentralization of Wage
Bargaining”, IZA Discussion Paper No. 6176, Bonn.
Davis, S.J., R.J. Faberman and J. Haltiwanger (2006), “The Flow Approach to Labor Markets: New Data
Sources and Micro–Macro Links”, Journal of Economics Perspectives, Vol. 20, No. 3, pp. 3-26.
de Cea, P., and J. J. Dolado (2013), Output Growth Thresholds for Job Creation and Unemployment
Reduction in Spain, Universidad Carlos III, mimeo,
http://www.eco.uc3m.es/temp/dolado2/GDP%20Growth%20Thresholds%20_PdC-JD_.pdf
de la Puebla-Pinilla, A. (2013), Las Paradojas del Despido Colectivo tras la Supresión de la Autorización
Administrativa, Teoria y Derecho, No. 13, pp. 129-156.
Dolado, J., S. Ortigueira and R. Stucchi (2012), “Does Dual Employment Protection Affect TFP? Evidence
from Spanish Manufacturing Firms”, CEPR Discussion Paper, No. 8763, London.
Druant, M., Fabiani, S., Kezdi, G., Lamo, A., Martins, F. and R. Sabbatini (2009), “How are firms’ wages
and prices linked: survey evidence in Europe”, ECB Working Paper Series No. 1084, European
Central Bank, Frankfurt.
Du Caju, P., Gautier, E., Momferatou, D. and M. Ward-Warmedinger (2008), “Institutional features of
wage bargaining in 23 European countries, the US and Japan”, ECB Working Paper Series No. 974,
European Central Bank, Frankfurt.
Düll, N. (2013), Collective wage agreement and minimum wage in Germany, Ad hoc request, European
Employment Observatory, http://www.eu-employment-observatory.net/resources/reports/1-
Germany-NationalAdHocResponseMinimumWage-final.pdf
Düll, N., P. Tergeist, U. Bazant and S. Cimper (2010), Activation Policies in Switzerland, OECD Social,
Employment and Migration Working Paper, No. 112, OECD Publishing, Paris,
http://dx.doi.org/10.1787/5km4hd7r28f6-en
European Central Bank (2010), Wage Dynamics in Europe: Final Report of the Wage Dynamics Network
(WDN), European Central Bank, Frankfurt.
Fabiani, S., K. Galuscak, C. Kwapil, A. Lamo, M., and T. Rõõm (2010) Wage Rigidities and Labor Market
Adjustment in Europe, Journal of the European Economic Association, Vol. 8(2-3), pp. 497-505.
Flanagan, R.J. (1999), “Macroeconomic Performance and Collective Bargaining: a Macroeconomic
Perspective”, Journal of Economic Literature, Vol. 37, pp. 1150-1175.
García-Pérez, J. I. (2008), “La Muestra Continua de Vidas Laborales: Una guía de uso para el análisis de
transiciones”, Revista de Economía Aplicada, vol. XVI (1), pp. 5-28.
García-Pérez, J. I., and F. Muñoz-Bullón (2011), “Transitions into Permanent Employment in Spain: An
Empirical Analysis for YoungWorkers”, British Journal of Industrial Relations, vol. 49(1), pp. 103-
143.
![Page 50: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/50.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 49
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Garcia-Serrano, C. and M. A. Malo (2013), Beyond the contract type segmentation in Spain: country case
study on labour market segmentation, Employment Working Paper No. 143, International Labour
Office, Geneva.
Golan, A., J. Lane and E. McEntarfer (2006), “The Dynamics of Worker Reallocation within and across
Industries”, Economica, Vol. 74, pp. 1-20.
Gomez-Abelleira, F. (2012), The Spanish Labour Reform and the Courts: Employment Adjustment and the
Search for Legal Certainty, Spanish Labour Law and Employment Relations Journal (April-
November 2012), Vol. 1, No. 1-2, pp. 31-46
Güell, M. and B. Petrongolo (2007), “How Binding Are Legal Limits? Transitions from Temporary to
Permanent Work in Spain”, Labour Economics, Vol. 14, pp. 153-183.
Gürtzgen, N. (2010), “Rent-Sharing and Collective Wage Contracts–Evidence from German
Establishment-Level Data”, Applied Economics, Vol. 42, No. 22, pp. 2835-2854.
Haipeter, T. and S. Lehndorff (2009), “Collective Bargaining on Employment”, Industrial and
Employment Relations Department Working Paper No. 3, International Labour Office, Geneva.
Haucap, J., and C. Wey (2004), "Unionisation Structures and Innovation Incentives", Economic Journal,
Vol. 114, pp. C140-C165.
ILO Committee of Experts on the Application of Conventions and Recommendations (2012), “Comments
on the ILO Termination of Employment Convention, 1982, No. 158”, International Labour Office,
Geneva.
IMF (2013), Spain 2013 Article IV Consultation, IMF Country Report No. 13/244, International Monetary
Fund ,Washington, D.C.
Izquierdo, M., Lacuesta, A. and S. Puente (2013), La reforma laboral de 2012: un primer análisis de
algunos de sus efectos, Boletín Ecónomico, Banco de España, Septiembre 2013.
Jahn, E., R. Riphahn and C. Schnabel (2012), “Feature: Flexible Forms of Employment: Boon and Bane”,
Economic Journal, vol. 122(562), pages F115-F124.
Jimeno, J.F. and C.Thomas (2013), “Collective bargaining, firm heterogeneity and unemployment”,
European Economic Review, Vol. 59, pp. 63–79.
Keune, M. (2010), “Derogation Clauses on Wages in Sectoral Collective Agreements in Seven European
Countries”, Eurofound, Dublin, www.eurofound.europa.eu/pubdocs/2010/87/en/2/EF1087EN.pdf.
La Caixa (2013), The deleveraging of the periphery countries, Monthly Report – Economic and Financial
Outlook, Number 372, October, http://www.pdf.lacaixa.comunicacions.com/im/eng/201310_eng.pdf
Lepage-Saucier, N., J. Schleich and E.Wasmer (2013), “Moving Towards a Single Labour Contract: Pros,
Cons and Mixed Feelings”, OECD Economics Department Working Papers, No. 1026, OECD
Publishing, Paris, http://dx.doi.org/10.1787/5k4c0vvc4zxv-en.
Mercader-Uguina, J., and A. de la Puebla-Pinilla (2013), Los Procedimientos de Despido Colectivo,
Suspensión de Contratos y Reducción de Jornada, Tirant Lo Blanch, Valencia.
![Page 51: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/51.jpg)
50 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Ministerio de Empleo y Seguridad Social (2013), Report evaluating the impact of the Labour Reform,
Madrid,
http://www.empleo.gob.es/es/destacados/HOME/impacto_reforma_laboral/Report_evaluating_the_i
mpact_of_the_Labour_Reform.pdf
Ministero del Lavoro e delle Politiche Sociali (2012), Rapporto Annuale sulle Comunicazioni Obbligatorie
2012, Rome, http://www.cliclavoro.gov.it/Barometro-Del-
Lavoro/Documents/Rapporto_CO/Rapporto_annuale_co.pdf .
Nicoletti, G., and S. Scarpetta (2005), “Product Market Reforms and Employment in OECD Countries”,
OECD Economics Department Working Paper, No. 472, OECD Publishing, Paris.
OECD (2013a), Employment Outlook, OECD Publishing, Paris, http://dx.doi.org/10.1787/empl_outlook-
2013-en
OECD (2013b), Detailed description of employment protection legislation, 2012-2013, Employment
Protection Database, http://www.oecd.org/els/emp/All.pdf.
OECD (2012), Employment Outlook, OECD Publishing, Paris, http://dx.doi.org/10.1787/empl_outlook-
2012-en .
OECD (2010), Employment Outlook, OECD Publishing, Paris, http://dx.doi.org/10.1787/empl_outlook-
2010-en .
OECD (2009), Employment Outlook, OECD Publishing, Paris, http://dx.doi.org/10.1787/empl_outlook-
2009-en .
OECD (2006), Employment Outlook, OECD Publishing, Paris, http://dx.doi.org/10.1787/empl_outlook-
2006-en.
OECD (2004), Employment Outlook, OECD Publishing, Paris, http://dx.doi.org/10.1787/empl_outlook-
2004-en.
OECD (1999), Employment Outlook, OECD Publishing, Paris, http://dx.doi.org/10.1787/empl_outlook-
1999-en.
OECD (1994), The OECD Jobs Study: Evidence and Explanations, Part II, OECD Publishing, Paris.
Palomo-Balda, E. (2013), “111 sentencias en materia de despido colectivo (y 2 más): un balance
provisional”, paper presented at the conference “Doctrina judicial sobre despidos colectivos”,
Madrid, May 2013, updated July 2013.
Paraire, X. (2012), “Les mouvements de main-d’oeuvre en 2011 : Une rotation élevée dans le tertiaire”,
Dares Analyses-Dares Indicateurs, No. 2012-071.
Puente, S. and P. Font (2013), “La evolución del empleo y del paro en el tercer trimestre de 2013, según la
Encuesta de Población Activa”, Boletín Ecónomico, Banco de España, Noviembre 2013.
Rusinek, M. and F. Rycx (2013), "Rent-Sharing under Different Bargaining Regimes: Evidence from
Linked Employer–Employee Data," British Journal of Industrial Relations, Vol. 51(1), pp. 28-58.
![Page 52: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/52.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 51
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
van Klaveren, M. (2011), “Sector-level bargaining and possibilities for deviations at company level –
Ireland”, Eurofound, Dublin,
http://www.eurofound.europa.eu/pubdocs/2010/875/en/1/EF10875EN.pdf
Visser, J. (2004), “Recent Trends and Persistent Variations in Europe’s Industrial Relations”,
www.csmb.unimore.it/on-line/Home/Prova/documento36007711.html.
![Page 53: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/53.jpg)
![Page 54: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/54.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 53
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
ANNEX
EMPIRICAL FRAMEWORK AND DETAILED ESTIMATION RESULTS
The evaluation of the 2012 Spanish labour market reform is an arduous task. First, the exercise is
performed shortly after the reform. Available data cover at best 18 months in the post-reform period, but
sometimes much less. Second, the inclusion of a large number of provisions, sometimes explicitly targeted
at different groups, does not allow the identification of a control group. Third, the reform also occurred in
the middle of a double-dip recession, at a time when the economic juncture restarted to deteriorate after a
short recovery that was so hesitant that never stopped employment levels from falling. This suggests that
one of the few available methods to analyse the role of the reform in the recent evolution of the Spanish
labour market is through regression-discontinuity models in which the effect of the reform is identified
through discontinuous patterns occurring at the time of its enforcement and the business-cycle is modelled
through observable controls and non-linear time trends.
The estimation strategy followed in this report identifies the joint effect of all the provisions included
in the reform by comparing labour market performance before and after February 2012. The key
identification assumption is that, conditional on control variables included in the model, labour market
performance evolves in a relatively smooth way, so that any discontinuous jump in performance
(conditional on control variables) can be attributed to the labour market reform (and other institutional
changes occurring simultaneously). The general regression-discontinuity model, which is estimated on
various sets of either quarterly or monthly data, can be written as:
tt
s
s
Rts
s
s
sRtittit DRtIRtIXYP
5
1
5
1
)()( (1)
where P is a performance variable which can be measured at time t either at the aggregate level or at a
more disaggregate level, indexed by i (such as the industry, the establishment or the individual level).
Y and X are vectors of aggregate and disaggregate (if relevant) confounding factors, respectively, notably
capturing composition effects, while R is the date of the reform,58 I is the indicator function (which, in eq.
(1) indicates the post-reform dummy), D stands for seasonal (e.g. quarterly or monthly) dummies, and
Greek letters stand for parameters to be estimated, except for , which represents a standard error term. The
parameter of interest is . A significant estimate for this parameter suggests a significant impact of the
reform. As suggested by Card and Lee (2008), since the identification is based on a time discontinuity,
standard errors are always adjusted for clustering at least on time (and, where relevant, also on other
dimensions). Moreover, probability weights are also used in order to ensure the relevance of results for the
dynamics of the aggregate business sector.
Given the lack of control group and the particular time at which the reform took place (see above),
properly isolating the effect of the reform from that of the business-cycle is key for the validity of the
smoothness assumption. In order to capture economic fluctuations, the baseline estimation models control
for the standardised unemployment rate, changes in employment registered with the general social security
regime (at the regional or industry level when microdata are used) and, most importantly, polynomial time
trends up to the 5th order. Baseline specifications include a polynomial trend up to the highest order n such
58. As the reform entered into force on February 12th, 2012, R is set at the beginning of 2012 in baseline
specifications with quarterly data and at the beginning of March 2012 in those on monthly data (except if
differently specified).
![Page 55: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/55.jpg)
54 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
that the n-th term is not insignificant. However, except when specified otherwise, the results are robust
both to changes in the order of the polynomial and to its exclusion from the specification. When a
sufficiently long number of periods is available after the reform, polynomial trends are alternatively
included as either homogeneous (same parameters before and after the reform) or heterogeneous (different
parameters before and after the reform), the latter represented by the interaction term between I and the
time trend in eq. (1). As a robustness check, employment changes are replaced with other – perhaps less
endogenous – controls for the business cycle, such as changes in either industry-level value added or
industry-level productivity growth (from quarterly national accounts) or the aggregate FEDEA index of
economic activity. The FEDEA index is an indicator of economic activity in Spain that measures the
economic cycle by using different sources of relevant information (GDP, Economic Sentiment Indicator,
industrial production, car sales, electricity consumption and workers affiliated in social security)59.
Obviously, misspecification of the empirical model might lead to identify a discontinuous shift in
performance around the date of a reform even when this shift occurs before the reform (and cannot
therefore be attributed to it). To validate the empirical model, therefore, placebo tests need to be run. These
tests consist in setting in the empirical model an earlier date than the actual reform date to estimate
possible discontinuities. In practice, this means setting an earlier date R – than the actual reform date – in
eq. (1) and re-estimating the equation. If discontinuous shifts in performance are really induced by the
reform, then no effect should be found at these anticipated dates. This is indeed the case for all the results
presented here, where placebo tests are run by anticipating the date of the reform by up to 3 quarters, in
models with quarterly data, or by up to 6 months in models with monthly data.
A second issue concerns possible manipulations around the threshold. For example, if the introduction
of the contrato emprendedores were anticipated, employers eligible for the subsidy could delay hiring from
before to after the reform in order to enjoy the subsidy. However, the details and the breadth of the reform
were never mentioned in the programme of the PPE before the November 2011 elections that the party won
and was not made public until well after the inaugural address of Prime Minister Rajoy made in front of the
parliament at the end of December 2011. It is therefore reasonable to assume that if threshold manipulation
occurred, that is if firms postponed certain choices until the approval of the reform, this phenomenon
concerned, at worst, only the period January-March 2012. Under these assumptions, threshold manipulation
is not an issue in the regressions estimated on quarterly data because the outcome of any firm choice, which
was delayed from January to March 2012, is included in the same quarter, with no effect on the average
outcome of the quarter. In the case of estimates based on monthly data, baseline models are re-estimated
excluding the period from January to March 2012 from the sample as an additional robustness check.
The fact that discontinuous patterns of labour market performance occurring around February 2012
are used to identify the effect of the 2012 reform makes it impossible to distinguish its impact from that of
other institutional changes occurring around the same date. In particular, this is the case of the Agreement
on Employment and Collective Bargaining 2012-2014, signed by the main business associations and trade
unions _ CEOE, CEPYME, CCOO and UGT - at the end of January 2012. This agreement promotes
internal flexibility, as a mean to preserve employment, as well as commits to wage moderation and the
expansion of performance-related pay.
A number of provisions, however, also either entered into force, de jure or de facto, after the date of
approval of the main reform (February 2012) or were modified since then. This makes more difficult to
identify unambiguously discontinuous patterns in the data. However, to the extent that the effects of
subsequent policy adjustments are likely to go in the same direction of those of the February 2012 reform,
results presented here are likely to represent only an underestimate of the true effect. Of course, if the
direction of the impact of subsequent policy interventions were different, this argument would be invalid.
59. For further information and methodology, see www.fedea.net/indice/indice-f.html.
![Page 56: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/56.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 55
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
This annex presents in detail methodology and empirical results dataset-by-dataset. The first section
presents the analysis of the impact of the reform on labour costs, based on data from Spanish Quarterly
Labour Cost Survey; the second section looks at hiring rates on the basis of data from the Encuesta de la
Populacion Activa; the third section studies new permanent contracts using the monthly statistics on
contracts published by SEPE; the fourth examines the differential trends in collective and individual
dismissals using the monthly statistics on unemployment benefits published by SEPE; the fifth studies
separation rates using the Encuesta de Coyuntura Laboral; and the sixth section examines transitions from
unemployment to employment using longitudinal data from the social security register.
Labour costs
Methodology
Unit labour costs measure the average nominal cost of labour per unit of output and are calculated as
the ratio of total nominal labour costs to real output. Of particular interest for this report are wage and
salary costs: the 2012 reform of collective bargaining is, in fact, expected to reduce the growth rate of
nominal wage and salary costs per unit of output in times of crisis. Obviously the evolution of labour costs
is strongly affected by the sectoral composition of activities. For this reason the analysis is carried out
using data at the industry level (at the 1 digit of the NACE rev. 2 classification), using data from the
Spanish Quarterly Labour Cost Survey and Quarterly National Accounts from the first quarter of 2005 to
the second quarter of 2013 and focussing mainly on the non-agricultural business sector. However, to the
extent that productivity data are not available at the same level of disaggregation (the industry
classification used in the quarterly national accounts being slightly more aggregate), the wage and salary
component of the harmonised index of labour costs per effective hour worked is used as dependent
variable and hourly labour productivity (and/or other variables capturing it) are included as controls. One
additional complication stems from seasonality: insofar as labour compensation vary significantly from
one quarter to another due to the timing of payment of specific bonuses, only year-on-year changes in the
harmonised index of labour costs can be meaningfully analysed. The estimated model, derived from the
general one of eq. (1), can be written as
itjt
s
Rt
s
jsjttcjttujt DDIRtXCPIuLC
5
1
)(logloglog
where LC is the labour cost index in industry j at time t (wage costs per effective hour, adjusted for
calendar days but not for seasonality), π is hourly productivity, CPI is the harmonised price index, X is a
vector of workforce and job characteristics (the share of open-ended contracts, the share workers with less
than one year of tenure, the share of three education categories, the share of four age categories, and the
share of women),60 u stands for the national-level standardised unemployment rate, R is the date of the
reform, I is the indicator function, D stands for quarterly and industry dummies, and denotes year-on-
year changes (t/t-4). As residuals in this model are inevitably serially-correlated, errors are adjusted for
clustering on both time and industry. As the reform is expected to affect the quarterly growth rate of labour
costs from the first quarter of 2012, it should affect their year-on-year growth only slowly. Therefore, in
contrast with other estimation exercises reported in this annex (see below), R is set at the beginning of the
second quarter of 2012 in the baseline specification. Finally, the time trend is included in two alternative
ways: in one series of specifications a 5th-order aggregate, homogeneous polynomial trend is included,61
while in a second series a linear trend is allowed to vary across industries to capture industry heterogeneity
in growth patterns. To the extent that only 5 quarters are observed where I is equal 1, the parameters of
polynomial trends are not assumed to change over time.
60. These data are from the labour force survey (Encuesta de la Populacion Activa, EPA).
61. The 5th order term is significant in all specifications.
![Page 57: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/57.jpg)
56 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Results
Table A1 presents the key results. In the two baseline models, the impact of the reform on the growth
of labour costs, conditional on productivity growth and other controls is estimated between 1.2 and
1.9 percentage points, significant at the 5% statistical level. Similar results are obtained with the following
alternative specifications (not reported in the table but available upon request): i) reducing the order of the
polynomial trend (or not including any trend); ii) substituting changes in industry-level log productivity
with changes in either aggregate productivity or industry-level log employment, or including the latter as
an additional covariate;62 and iii) estimating the models with seasonally-adjusted data.
Table A1. The effect of the reform on year growth of labour costs
Industry-level equations, quarterly data
Post-reform dummy Placebo tests
Baseline
One quarter before
Two quarters before
Three quarters before
(1) (2) (3) (4)
Model:
5-th order polynomial -0.019** -0.006 0.020* 0.021
(0.009) (0.011) (0.011) (0.013)
Industry-specific trend -0.012** -0.005 0.004 0.009*
(0.005) (0.006) (0.005) (0.005)
Note: The dependent variable is the year-on-year quarterly change of the logarithm of the industry-level wage and salary component of the harmonised labour cost index. Each cell refers to a different specification. Robust standard errors, clustered on time and industry, in parentheses. All equations also control for the unemployment rate, change in log hourly productivity, quarter and industry dummies, changes in the log consumer price index, and changes in the industry composition of the workforce (that is, for each industry, the share of: open-ended contracts, workers with less than one year of tenure, 3 education categories, 4 age categories, and women). The post-reform dummy takes value 1 from the second quarter of 2012. In placebo tests it is replaced with a dummy taking value 1 from the fourth, third and second quarter of 2011 in columns 2, 3 and 4, respectively. ** p<0.05, * p<0.1
Placebo tests are implemented by replacing the post-reform dummy with dummies taking value 1
from the beginning of either the fourth or the third or the second quarter of 2011.63 In all cases, estimates
show either an insignificant or a positive and marginally significant effect of the “placebo”, therefore
confirming that the estimated effects estimated in baseline models can be attributed to the reform.
Hiring rates
Methodology
Hiring rates are normally defined as new hires in a given period over employment in the same period.
The microeconomic equivalent of this concept at the individual level is the probability of having been
hired in a period of duration h before the reference week. In other words, one can estimate the effect of the
62. Short-term changes in employment and productivity are typically very strongly negatively correlated, therefore
changes in the level of employment can be used to capture variations in productivity, since productivity by
industry is not available at a sufficiently disaggregate level (the expected sign of the coefficient of employment
is therefore negative). This allows also extending the number of covered industries. In fact, mining, water and
electricity and gas are excluded from baseline specifications since productivity data are not available for these
industries. When employment is also included in these specifications while already controlling for productivity
changes, this is done to better control for possible worker selection.
63. Since the first quarter of 2012 is the quarter when the reform was implemented, a placebo test setting the
placebo-reform dummy equal to 1 since the start of that quarter cannot be easily interpreted and is therefore not
reported.
![Page 58: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/58.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 57
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
reform on the hiring rate by estimating its impact on the probability that an employee has tenure shorter
than h. The following regression discontinuity model is estimated on individual quarterly data from the
labour force survey (Encuesta de la Populacion Activa, EPA, cross-sectional file with a 2-digit industry
classification) from the first quarter of 2008 to the second quarter of 2013:64
ijttRt
s
s
sijtjtetuijt DIRtXEuH
5
1
)(loglog (2)
where H is an indicator variable taking value 1 if the individual i in region j has job tenure lower or equal
to 3 months at time t. H refers alternatively to all hires, hires on permanent contracts and hires on
temporary contracts. The sample is restricted to all wage and salary employees in the non-agricultural
business sector. As far as control variables are concerned, u is the quarterly standardised unemployment
rate, E is regional employment, X is a vector of individual characteristics (70 2-digit industry dummies,
gender, 10 age categories, 7 education categories, 10 occupational categories, and dummies for being non-
native, having multiple jobs and having changed municipality in the last year).65 Finally, as in eq. (1), R is
the date of the reform (set at the first quarter of 2012), I is the indicator function, D stands for quarterly
dummies, denotes changes and Greek letters stand for parameters to be estimated, except for , which
represents a standard error term. Finally, a 4th-order aggregate homogeneous polynomial trend is
included,66 in baseline specifications. To the extent that only 6 quarters are in the post-reform sample, the
parameters of the polynomial trends are not assumed to change over time.
Results
Baseline estimates, reported in Table A2 suggest that the 2012 reform increased the hiring rate –
measured as the proportion of employees with 3 month of tenure or less – by about 8% (that is a
percentage-point marginal effect of 0.64 percentage points). With a 13% increase (corresponding to a
percentage-point marginal effect of 0.14 percentage points), the estimated percentage effect is larger in the
case of hiring on open-ended contracts. Very similar results – in percentage terms – are obtained if the
sample is restricted to employees with an open-ended contract only. A somewhat stronger effect is
estimated in the case of hiring on open-ended full-time contracts – 18% increase – corresponding to a
percentage-point marginal effect of 0.16 percentage points). By contrast, the estimated percentage effect is
much smaller in the case of hiring on temporary contracts, with a 7% increase (corresponding to a
percentage-point marginal effect of 0.5 percentage points).
All the results presented in Table A2 are robust to i) specifying eq. (2) as a probit model instead of a
linear probability model; ii) increasing the order of the polynomial trend67 (or not including any trend);68
iii) substituting changes in regional employment with, alternatively, changes in industry real value added
or changes in unemployment rates; and iv) excluding all aggregate controls except polynomial time trends
from the specification. By contrast, the only exception is the result on hiring on fixed-term contract, which
is not robust to the exclusion of the polynomial trend.
64. The sample is limited to this period due to the change in industry classification that occurred between 2007
and 2008.
65. All individual data are from EPA.
66. The 4-th order term is significant in all specifications.
67. Although higher-order terms turn out to be statistically insignificant.
68. However, if a lower-order polynomial trend is included, the effect of the reform on the hiring rate of
permanent workers is insignificant, although still positive.
![Page 59: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/59.jpg)
58 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Table A2. The effect of the reform on the probability of having tenure shorter than three months
Individual-level equations, quarterly data
(1) (2) (3) (4) (5)
Dependent variable Hired Hired,
permanent Hired,
permanent Hired, full-time,
Permanent Hired,
fixed-term
Post-reform dummy 0.00646*** 0.00135** 0.00183** 0.00156** 0.00511**
(0.00185) (0.00061) (0.00080) (0.00062) (0.00192)
Sample restricted to perm. workers No No Yes No No
Observations 712,393 712,393 551,429 712,393 712,393
R-squared 0.066 0.004 0.012 0.004 0.067
Note: Hired stands for a dummy indicating that the worker has been hired in the three months preceding the reference week. Robust standard errors, clustered on time, in parentheses. All equations control for the unemployment rate, change in regional employment, quarterly dummies, a 4
th order polynomial in time (quarters), 70 2-digit industry dummies, gender, 10 age categories, 7 education
categories, 10 occupational categories, and dummies for being non-native, having multiple jobs and having changed municipality in the last year. *** p<0.01, ** p<0.05.
Placebo tests are then run for all specifications by “anticipating” the date of the reform by up to
3 quarters (Table A3). None of the placebo effect is significantly positive. However for both total hiring
and, in particular, hiring on temporary contracts, placebo effects are significantly negative. This suggests
some possible misspecification of the business-cycle component, so that the attribution of the increase in
hiring on temporary contract to a genuine effect of the reform must be considered more uncertain.
Table A3. The effect of different placebos on the probability of having tenure shorter than three months
Individual-level equations, quarterly data
(1) (2) (3) (4) (5)
Dependent variable Hired Hired,
permanent Hired,
permanent Hired, full-time,
Permanent Hired,
fixed-term
Placebo:
1 quarter before -0.00073 0.00028 0.00041 0.00063 -0.00101
(0.00332) (0.00086) (0.00114) (0.00084) (0.00293)
2 quarters before -0.00447** 0.00028 0.00028 0.00058 -0.00474***
(0.00200) (0.00078) (0.00102) (0.00074) (0.00140)
3 quarters before -0.00615*** -0.00142* -0.00192* -0.00122 -0.00473***
(0.00156) (0.00077) (0.00104) (0.00086) (0.00152)
Sample restricted to perm. workers No No Yes No No
Note: Hired stands for a dummy indicating that the worker has been hired in the three months preceding the reference week. Each cell refers to a different specification. Robust standard errors, clustered on time, in parentheses. All equations control for the unemployment rate, change in regional employment, quarterly dummies, a 4
th order polynomial in time (quarters), 70 2-digit industry
dummies, gender, 10 age categories, 7 education categories, 10 occupational categories, and dummies for being non-native, having multiple jobs and having changed municipality in the last year. *** p<0.01, ** p<0.05, * p<0.1
New permanent contracts
The monthly statistics on contracts, published by SEPE, allows replicating and deepening the analysis
of hiring on permanent contracts. In fact, while EPA covers a sample of few tenth of thousands employees
and is available only at the quarterly base, contract statistics are published monthly and are based on
compulsory administrative declarations from all employers and therefore cover the entire population. In
![Page 60: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/60.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 59
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
addition a breakdown of new contracts is available by firm size, which is important given that certain
provisions of the 2012 reform are applicable only to firms below a certain size threshold.
Methodology
In order to assess the impact of the reform on hiring on permanent contract, the following class of
regression-discontinuity models is estimated, using on aggregate monthly data for the period January 2006-
July 2013:
tt
s
s
RtsRt
s
s
sttdtut DRtIIRtXuuNP
5
1
5
1
)()(log (3)
where NP is the number of new permanent contracts at time t, u is the monthly standardised unemployment
rate, X is a vector of variables capturing workforce composition of new hires (shares of 3 education
categories, 5 age categories and women), R is the date of the reform, I is the indicator function, D stands
for monthly dummies, denotes changes and Greek letters stand for parameters to be estimated. All data
are from the monthly statistics on contracts published by SEPE, except for the unemployment rate. As the
identification is based on a time discontinuity, standard errors are adjusted for clustering on time. Finally, a
5th-order aggregate polynomial trend is included, in baseline specifications.69 Since 18 observations are
available in the post-reform period, the time trend is alternatively specified as homogeneous (same
parameters before and after the reform) and heterogeneous (different parameters before and after the
reform).
Eq. (3) is estimated for the whole economy and then separately for firms above and below the threshold
of 50 employees (since only firms below the threshold are eligible for the contrato emprendedores introduced
by the reform). Then, both groups of firms (above and below this threshold) are divided further into those
with size above and below 25 employees and those of size above and below 100 employees.
Results
As shown in Table A4, the 2012 reform increased hiring on regular open-ended contracts by
20%-30%. If firms are split in two subsamples – those with 50 or fewer employees and those with more
than 50, no significant increase is observed for the largest firms. By contrast, no matter how the time trend
is specified, the reform is estimated to have increased the number of new contracts by about 42%, and this
estimated impact is always significant at the 1% level. For these two groups of firms the predicted effect of
the reform is plotted in Figures A1 and A2. For firms with 50 or fewer employees, the baseline model with
homogeneous time trend predicts that the reform raised the number of new open-ended contracts by at least
25 000 units, while no effect is detectable for larger firms. These trends are also reflected in the evolution
of the share of regular open-ended contracts in total hiring, which is estimated to have been raised by about
3 percentage points (or about 30%) in firms with 50 or fewer employees, while no significant effect
emerges for larger firms.70
69. The 5-th order term is significant in all specifications.
70. These figures (not shown in the table) are obtained by replacing the log of the number of new contracts
with the share of permanent contracts in new hires in eq. (3). All results presented here are robust if the
share of regular open-ended contracts in total hiring is used as dependent variable instead of the number of
open-ended contracts.
![Page 61: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/61.jpg)
60 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Table A4. The effect of the reform on new permanent contracts
Panel A. Homogeneous trend, monthly data
(1) (2) (3) (4) (5) (6) (7)
Sample (firm size) All 1to50 >50 1to25 26to50 51to100 >100
Post-reform dummy 0.217* 0.421*** 0.132 0.463*** 0.281*** 0.231** 0.067
(0.124) (0.110) (0.084) (0.109) (0.095) (0.103) (0.080)
Observations 91 91 91 91 91 91 91
R-squared 0.957 0.949 0.970 0.946 0.959 0.953 0.972
Panel B. Heterogeneous trend
(1) (2) (3) (4) (5) (6) (7)
Sample (firm size) All 1to50 >50 1to25 26to50 51to100 >100
Post-reform dummy 0.334*** 0.424*** 0.150 0.456*** 0.266* 0.286** 0.099
(0.108) (0.101) (0.116) (0.0942) (0.136) (0.117) (0.127)
Observations 91 91 91 91 91 91 91
R-squared 0.965 0.959 0.971 0.956 0.966 0.960 0.973
Note: Dependent variable: logarithm of new open-ended contracts. Robust standard errors, clustered on time, in parentheses. Each equation includes level and changes in the standardised unemployment rate, the shares of 3 education categories, 5 age categories and women in new contracts and a 5-th order polynomial trend in time (months), which in Panel B is assumed to differ between before and after the reform. *** p<0.01, ** p<0.05, * p<0.1
When the data are further disaggregated by firm size, the greatest estimated expansion in the number
of new permanent contracts induced by the reform is found in very small firms with 25 employees or less
(45%-46%; see Table A4). In firms between 26 and 100 employees the effect is somewhat more limited.
The reform is estimated to have raised the number of new permanent contracts by 26%-28% in firms with
26 to 50 employees and by 23%-28% in firms with 51 to 100 employees. By contrast, no significant effect
is observed for larger firms.71
All the results presented in Table A4 are robust to i) changing the order of the polynomial trend (or
not including any trend); ii) excluding workforce controls; iii) substituting polynomial trends with year
dummies; iv) substituting changes in the unemployment rate with either changes in the FEDEA index or
the logarithm of employment as measured by the number of workers enrolled in the general social security
regime; and v) excluding all aggregate controls except polynomial time trends from the specification.
Placebo tests are then run for all specifications by “anticipating” the date of the reform by up to
6 months. Table A5 presents the results for the case of homogenous time trends. Results are however
similar in the case of heterogeneous trends and/or if changes in the unemployment rate are replaced by
71. No matter how the time trend is specified, the effect on firms with 25 employees or less appears also
significantly larger than that of firms with 26 to 50 employees. Similarly the effect on firms with more than
100 employees appears significantly smaller than that of 51 to 100 employees. By contrast, the difference
in the estimated impact between firms with 26 to 50 employees and 51 to 100 employees is statistically
insignificant at conventional levels.
![Page 62: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/62.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 61
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
changes in the FEDEA index. In all cases where a significant effect of the reform is estimated with eq. (3),
placebo tests show no significant effect if the date of the reform is set at the beginning of January 2012 or
before. This suggests that the estimated effects in Table A4 can be genuinely attributed to the reform.
Figure A1. New hires with a permanent contract in firms with 50 or fewer employees, 2006-2013
Monthly hires, observed and fitted by the baseline model
Note: The figure presents the number of new permanent contracts signed each month, as observed in raw data and as fitted using estimates of Table A4, Column 2 (including netting out the effect of the post-reform dummy).
Figure A2. New hires with a permanent contract in firms with more than 50 employees, 2006-2013
Monthly hires, observed and fitted by the baseline model
Note: The figure presents the number of new permanent contracts signed each month, as observed in raw data and as fitted using estimates of Table A4, Column 3 (including netting out the effect of the post-reform dummy).
0
20000
40000
60000
80000
100000
120000
140000
160000
180000
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
2006 2007 2008 2009 2010 2011 2012 2013
Observed Fitted Fitted withouth the reform
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
100000
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
2006 2007 2008 2009 2010 2011 2012 2013
Observed Fitted Fitted withouth the reform
![Page 63: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/63.jpg)
62 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Table A5. The effect of different placebos on new permanent contracts
Homogeneous trend, monthly data
(1) (2) (3) (4) (5) (6) (7)
Sample (firm size) All 1to50 >50 1to25 26to50 51to100 >100
Placebo:
Placebo-reform Jan 12 -0.030 0.034 0.066 0.043 0.004 -0.008 0.070
(0.090) (0.110) (0.084) (0.112) (0.118) (0.113) (0.089)
Placebo-reform Dec 11 -0.061 -0.044 -0.049 -0.046 -0.104 -0.155 -0.016
(0.090) (0.101) (0.089) (0.099) (0.114) (0.105) (0.084)
Placebo-reform Aug 11 -0.119 -0.135 -0.123 -0.146* -0.135 -0.193* -0.061
(0.090) (0.084) (0.081) (0.079) (0.122) (0.107) (0.069)
Note: Dependent variable: logarithm of new open-ended contracts. Each cell refers to a different specification. Robust standard errors, clustered on time, in parentheses. Each equation includes level and changes in the standardised unemployment rate, the shares of 3 education categories, 5 age categories and women in new contracts and a 5-th order polynomial time trend, which in Panel B is assumed to differ between before and after the reform. * p<0.1
A breakdown by type of transition is also available in the data, which allows estimating eq. (3)
separately for new employment relationships and conversions of temporary contracts into open-ended ones
(Table A6).
Table A6. The effect of the reform on new permanent contracts, by type of transition
Panel A. Conversions
(1) (2) (3) (4) (5) (6) (7)
Sample (firm size) All 1to50 >50 1to25 26to50 51to100 >100
Post-reform dummy 0.159 0.306* 0.213* 0.325** 0.266* 0.387** 0.121
(0.182) (0.163) (0.123) (0.162) (0.136) (0.161) (0.102)
Observations 91 91 91 91 91 91 91
R-squared 0.925 0.924 0.921 0.919 0.931 0.907 0.924
Panel B. New employment relationships
(1) (2) (3) (4) (5) (6) (7)
Sample (firm size) All 1to50 >50 1to25 26to50 51to100 >100
Post-reform dummy 0.240*** 0.498*** 0.046 0.562*** 0.267*** 0.041 0.012
(0.090) (0.098) (0.075) (0.103) (0.058) (0.072) (0.083)
Observations 91 91 91 91 91 91 91
R-squared 0.972 0.963 0.982 0.957 0.976 0.975 0.979
Note: Dependent variable: logarithm of new open-ended contracts, by type, estimated on monthly data. Robust standard errors, clustered on time, in parentheses. Each equation includes level and changes in the standardised unemployment rate, the shares of 3 education categories, 5 age categories and women in new contracts and a 5-th order, homogeneous polynomial time trend, which in Panel B is assumed to differ between before and after the reform. *** p<0.01, ** p<0.05, * p<0.1
As far as conversions are concerned the effect of the reform is the greatest for firms with 51 to
100 employees (with a 39% estimated increase) followed by those with 1 to 25 employees (32% increase).
However, while for the latter the estimated increase in new open-ended employment relationships is much
![Page 64: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/64.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 63
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
larger (56% increase), there is essentially no effect on these for the former. As regards firms with 26 to
50 employees the effect of the reform appears similar on both conversions and new open-ended
employment relationships (about 27% in both cases), although less precisely estimated for the former.72
Collective and individual dismissals
Inflows onto unemployment benefits, by reason of inflow, are a reliable source of data to distinguish
the effect of the 2012 reform on individual and collective dismissals.73
Methodology
The same class of regression-discontinuity models used in the previous section for new contracts
(eq. 3) can be used to estimate the impact of the reform on the number of individual and collective
dismissals, using on aggregate monthly data (January 2007-July 2013). Three differences are however in
order: i) the dependent variable in this case will be replaced by either the logarithm of monthly inflows
onto unemployment benefits due to, alternatively, collective or individual dismissals or the share of
inflows due to collective dismissals in total inflows due to dismissals; ii) the logarithm of inflows due to
other reasons than that those measured in the dependent variable replaces changes in the unemployment
rate (since the former seems more appropriate in this case); and iii) to the extent that this model is used to
study mainly collective dismissals and at least one-month consultations with trade-unions are required by
post-reform regulations in the case of collective dismissals, the post-reform dummy takes value one only
one-month after the reform (therefore taking value 1 from April 2012).74 All data are from the monthly
statistics on unemployment benefits (Prestaciones por desempleo) published by SEPE. Inflows by
dismissal exclude those resulting from conciliation settlements, court rulings or contract suspensions, but
include those ex ley 45/2002 (individual dismissals for which benefits are claimed before that a court can
rule whether the dismissal is fair or unfair). Inflows data considered here are limited to workers who are
eligible to standard unemployment benefits because of sufficient previous contributions and contribution
periods (Prestaciones de nivel contributivo). They exclude, therefore, workers eligible only for other type
of unemployment-related subsidies (Prestaciones de nivel asistencial). Since published data do not contain
a breakdown of inflows by workers’ characteristics, these controls are not included. As the identification is
based on a time discontinuity, standard errors are adjusted for clustering on time.
In order to capture the business-cycle (over and above levels of the unemployment rate and other
inflows onto unemployment benefits), models of the type of eq. (3) include a polynomial trend up to
5th order, which can be either homogeneous (same parameters before and after the reform) or
heterogeneous (different parameters before and after the reform) depending on the specification. The
baseline specification includes a homogeneous polynomial trend up to the highest order n such that the n-th
term is not insignificant (in practice, the 2nd order in the case of the number of collective dismissals, the
3rd
order in the case of its share and the 5th order in the case of individual dismissals).
Results
Taking baseline estimates with homogeneous trends at face value, the 2012 reform is estimated to
have reduced inflows onto unemployment benefits for reasons of collective dismissals by about 32%
72. Similar estimates are obtained by using heterogeneous trends and/or replacing changes in unemployment
with changes in the FEDEA index.
73. However, it must be kept in mind that available data are limited to workers who have sufficiently long
contribution histories to be eligible for unemployment benefits.
74. This seems appropriate to minimise the risk of counting as subsequent to the reform dismissals that were
originated by a collective-dismissal procedure which in fact started before the reform.
![Page 65: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/65.jpg)
64 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
(Table A7). By contrast, individual dismissals decreased by about 6% only. Moreover, this effect is
significant only at the 10% level. Results are similar with heterogeneous trends (not shown in the table).
This evolution is reflected in the estimated effect of the share of collective dismissals in total dismissals
(see also Figure A3). In fact, the reform is estimated to have reduced this share by about 3.1 percentage
points (that is about 30%).
Table A7. The effect of the reform on inflows onto unemployment benefits due to dismissal
Monthly data, homogeneous trends
(1) (2) (3)
Dependent Variable Log Collective
dismissals Log Individual
dismissals Share Collective
dismissals
Post-reform dummy -0.316*** -0.060* -3.146***
(0.084) (0.035) (0.856)
Observations 79 79 79
R-squared 0.948 0.964 0.928
Note: Dependent variable: Columns 1 and 2: logarithm of inflows onto unemployment benefits for the reason indicated in the column title; Column 3: share of collective dismissals in inflows due to any type of dismissal. All equations control for the unemployment rate, a polynomial trend in time (months) and inflows in unemployment benefits for reasons different from those reported in the column title. A 2
nd-order polynomial is included in Column 1, a 5
th-order polynomial in Column 2 and 3
rd-order polynomial in Column 3. Robust
standard errors, clustered on time, in parentheses. *** p<0.01, ** p<0.05, * p<0.1
Figure A3. Share of inflows onto unemployment benefits due to collective dismissal in total inflows due to dismissal, 2007-2013
Monthly percentage share of inflows, observed and fitted by the baseline model
Note: The figure presents the monthly percentage share of inflows in unemployment benefits due to collective dismissal in total inflows due to dismissal, as observed in raw data and as fitted using estimates of Table A7, Column 3 (including netting out the effect of the post-reform dummy)..
0
5
10
15
20
25
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
Oct
Jan
Apr Ju
l
2007 2008 2009 2010 2011 2012 2013
Observed Fitted Fitted withouth the reform
![Page 66: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/66.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 65
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Results on collective dismissals are robust to i) substituting homogeneous trends with heterogeneous
trends; ii) changing the order of the polynomial trend (or not including any trend);75 and iii) replacing
changes in the unemployment rate with changes in the FEDEA index. By contrast, results on individual
dismissals are sensitive to the specification of the polynomial trend. In fact, if a 2nd or 3rd-order polynomial
is specified, the point estimate is even positive, although insignificant.
Placebo tests are then run for all specifications by “anticipating” the date of the reform by up to
6 months. Figure A4 presents the results for the share of collective dismissals. It turns out that if the the
date of the “placebo-reform” is set even one single month before, its estimated effect would be
insignificant, no matter whether the trend is modelled as homogeneous or heterogeneous. This suggests
that the estimated effects in Table A4 can be genuinely attributed to the reform.
Figure A4. The effect of different placebos on the share of inflows onto unemployment benefits due to collective dismissals
Panel A: Homogeneous trends
Panel B: Heterogeneous trends
Note: The figure report estimated coefficients and confidence interval obtained by estimating the specification reported in Table A7, Column 3, substituting the post-reform dummy with a dummy taking value 1 from the indicated dates. Apr-12 corresponds to the estimated effect of the true reform (measured through the coefficient of the post-reform dummy).
75. However, the effect on the share of collective dismissals become insignificant is no trend (or a linear,
homogeneous trend) is included.
-6
-5
-4
-3
-2
-1
0
1
2
95% lower bound Point estimate 95% upper bound
-6
-4
-2
0
2
4
6
95% lower bound Point estimate 95% upper bound
![Page 67: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/67.jpg)
66 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Worker separations at the establishment level
Methodology
Worker separations at the establishment level are analysed using the Encuesta de Coyuntura Laboral
(ECL), an establishment-level76 survey that follows establishments over time and registers its worker
flows. Around 12,800 establishments are surveyed each quarter. This sample is representative of all
establishments in the private sector and covers all workers registered in the General Regime of the Spanish
Social Security Register (as well as those in the Special Mining Regime).
Separation rates at the establishment level are defined as the number of separations in each quarter
divided by the average of total employment between the start and the end of the period at each
establishment. Separation rates are computed both for full-time and part-time employment, and
distinguishing between those separations that are due to end of contract for temporary workers, due to
dismissals of permanent workers, due to end of season for seasonal workers with an open-ended contract –
“fijos discontínuos” – and due to other reasons (including retirements, voluntary separations, deaths, etc.).77
The following regression discontinuity model is estimated on quarterly establishment-level data from
ECL for the period 2006 to 2012, for which micro data are available:
jtjtRt
s
s
sjtjttjt QIRtFEuSR
5
1
)(loglog (4)
where SR corresponds to the separation rate of the establishment j in quarter t. SR refers alternatively to
separation rates of all workers, of full-time/part-time as well as separation rates due to end of contract,
dismissals, end of season and other reasons. The sample is restricted to all establishments in the private
sector with at least ten workers on average for the whole period they are observed in the data. The
estimation has as cyclical controls the quarterly standardised unemployment rate at the national level (u)
and the change in employment level (in logs) at the regional level ( log E). In addition, several
establishment characteristics F are included: industry, establishment size, region as well as average age and
qualification structure in each sector in the region (extracted from LFS data). R is the date of the reform, I
is the indicator function, Q stands for quarterly dummies, and denotes changes. Greek letters stand for
parameters to be estimated, except for which represents a standard error term. The standardised
unemployment rate is from OECD labour market statistics, while regional employment is measured by the
number of workers enrolled in the general social security regime. Establishment fixed-effects are
included in fixed-effects regressions. As the identification is based on a time discontinuity, standard errors
are adjusted for clustering on time.
An additional variable is added to control for the share of permanent (and temporary) contracts in the
establishment, which is important when differentiating between types of separations. The results are robust
to the inclusion of this variable.
76. The unit of observation is the social security account number (Cuenta de Cotización), which is unique for
each firm at the provincial level. It does not correspond exactly to the plant or the firm level (as those firms
established in different provinces are not linked), although for a matter of simplicity we will refer to each
social security accounting unit as an “establishment ”, hereafter. For each Cuenta de Cotización, it is
possible to identify if the account corresponds to establishments with only one or several plants in the same
province. Almost 90% of the observations have only one plant.
77 . The separation rate of each type of workers (full/part-time, etc) is defined as the ratio between the number
of separations of each type of workers divided by the average of total employment between the start and
the end of the period at the establishment.
![Page 68: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/68.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 67
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
The estimation of equation (4) is performed using ordinary least squares as well as fixed-effect
models78 for all establishments in the private sector with more than 10 employees. The equation is also
estimated for small establishments with less than 50 employees only, with similar results. Results shown
include a polynomial time trend up to the 5th order but are robust to changing the order of the polynomial
time trend.
Results
The baseline results (see results in Table A8) suggest that the 2012 reform significantly reduced the
share of workers leaving the establishment in a given quarter. The estimated coefficient for the overall
separation rate implies that the reform reduced overall separation rates by about 24%, although the exact
point estimates differ slightly depending on the specification. The effect of the reform has been particularly
sizeable for part-time workers (for which separation rates are 41% lower due to the reform).
Table A8. The effect of the reform on separation rates at the establishment level
Quarterly data
Post-reform dummy Placebo tests
Baseline
One
quarter
before
reform
Two
quarters
before
reform
Three
quarters
before
reform
(1) (2) (3) (4)
Separation rates:
Overall -0.018*** -0.010* 0.005 0.006
(0.006) (0.006) (0.006) (0.005)
Full-time -0.006* -0.004 0 0.003
(0.003) (0.003) (0.004) (0.003)
Part-time -0.010** -0.007 0.008 0.005
(0.004) (0.005) (0.005) (0.004)
End of Contract -0.016** -0.016** 0.017** 0.013*
(0.006) (0.008) (0.008) (0.006)
Dismissal -0.002 -0.002** -0.001 0.002*
(0.001) (0.001) (0.001) (0.001)
End of Season -0.009** -0.008 0.013*** 0.009**
(0.004) (0.005) (0.004) (0.004)
Other 0.001 0.003** 0 -0.002
(0.002) (0.001) (0.002) (0.001)
Note: Each cell corresponds to the OLS coefficient of the post-reform (or the placebo test) dummy from a regression with each separation rate as dependent variable. All equations control for the unemployment rate, change in regional employment, a 5
th order
polynomial in time (quarters), establishment size, quarterly dummies, age and education composition of the establishment workforce and region and sector dummies. Robust standard errors, clustered on time in parentheses. ***, **, *: significant at the 1%, 5% and 10% level, respectively.
In addition, establishment-level data allows disaggregating flows by contract type and reason of
separation. The results point to a significant impact of the reform on separation rates due to the end of
contract for temporary workers (for which separation rates decreased by 31%) and for seasonal workers
78 . Alternatively, the estimation of equation (4) using a Tobit model is performed and similar results as OLS
are found. In addition, substituting separation rates by the log of the absolute value of separations does not
modify the results significantly.
![Page 69: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/69.jpg)
68 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
(71% decrease), but show no impact on dismissal rates and separations for other reasons. An impact of the
reform decreasing separation rates is still found when controlling for individual establishment effects,
although the effect is not significant (not shown in the table), suggesting that the reform affected separation
rates mainly by changing the composition of establishments.
Placebo tests are run for all the estimations by “anticipating” the date of the reform for up to
3 quarters. This is equivalent to substituting the date of the reform R by an earlier date and re-estimating
the equation (4). The coefficient of the placebo reform, when the its date is set just one quarter before the
true reform, is still negative and significant at the 5% for the separation rate due to the end of contract for
temporary workers, suggesting caution in the interpretation of the baseline estimates as regards this
covariate. Other placebo tests corresponding to two and three quarters before the reform show no
significant effect for all the estimations, except for a positive and significant one in the cases of end of
contract and end of season, which again suggests caution in interpreting the baseline results concerning
these variables.
An additional robustness test has been undertaken by replacing the change in log regional
employment ( log E) by the FEDEA index of economic activity. The results are not sensitive to using one
or the other indicator of the economic cycle.
Transitions from unemployment to employment
Methodology
Transitions from unemployment to employment are analysed using the Muestra Continua de Vidas
Laborales (MCVL). This longitudinal dataset from social security registers covers employment histories of
over one million individuals, making it a very good data source to study worker transitions out of
unemployment. The MCVL covers around one in twenty persons registered in the social security and is
representative of the whole population that had a relationship with social security in a given year79. These
data are used to study the impact of the reform on the hazard rate out of unemployment using a discrete-
time competing-risk duration model. The empirical strategy followed is similar to the one followed in
García-Pérez and Munoz-Bullón (2011).
The probability of exiting unemployment after a spell of duration d to enter either temporary
employment or permanent employment is simultaneously estimated for the two types of exits within a
competing-risk framework. Both possible transitions are modelled using a logistic distribution as given by:
))()ln(*logloglogPr
)ln(*)ln()ln()ln((
)|,Pr(
3
1
217
162514
3
3
2
210
tRt
s
s
sitjtjttit
ititititititit
ititit
mIRtUEEuevEmpl
UXXXUUUF
dULdU
(5)
where U refers to the duration of the unemployment spell, which ends at time t for the worker i, L is after-
transition employment contract-type (permanent or temporary) and 1X refers to the following individual
characteristics: age and education categories and an indicator of whether the individual receives
unemployment benefits or not. These are also included in interaction with the duration of the
unemployment spell. 2X stands for controls for gender, migrant status and region, while PrevEmpl for
characteristics of the previous job (sector, type of contract, and if the individual was dismissed or not) as
79. See García-Pérez (2008) for further details on the MCVL data and its usefulness to study labour market
transitions.
![Page 70: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/70.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 69
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
well as the number of times the individual has been unemployed and the percentage of time he has been
employed throughout his labour market career. Specifications include also as cyclical controls the monthly
unemployment rate at the national level (u) and the change in log employment at the provincial level
(logE). However, in order to capture the business-cycle (over and above control variables), models of the
type of eq. (5) include a polynomial trend up to the highest order n such that the n-th term is not
insignificant (in practice, the 3rd order or 5th order depending on the specification). Finally, R is the date of
the reform, I is the indicator function, m stands for monthly dummies, and denotes changes. Greek letters
stand for parameters to be estimated. In order to distinguish the estimated impact of the reform between
large and small establishments (to capture threshold effects) a four-fold competing-risk model is also
estimated, with four possible exiting states (permanent/temporary in large/small establishments).80
Equation (5) is estimated for a sample of Spanish workers aged 16-64 during the period 2003-2012
(2006-2012 for the subsample differentiating by establishment size). Transitions from unemployment to
employment are conditional on being unemployed for at least one day and durations are censored at
30 months of unemployment – that is, those spells lasting more than 30 months are considered to be
censored at the 30th month.
Results
The baseline results for the transitions from unemployment to employment (shown in Table A9)
indicate that the 2012 reform increased significantly the probability of leaving unemployment and entering
employment for any unemployment duration. The effect of the reform is significant as regards the
probability for the unemployed of being hired on a permanent contract, which has increased by 24%, on
average, during the first 6 months in unemployment where most transitions occur, although this probability
remains very small in absolute terms. The effect of the reform on the probability to exit from
unemployment into permanent employment is more sizeable for those being unemployed for less than
7 months than for those unemployed between 7 and 13 months or more than 13 months (Figure A5).
The reform seems to have increased as well the probability of leaving unemployment and entering
temporary employment, although the effect is not significant in most specifications.81 Thus, for
unemployment durations no longer than 6 months, the reform appears to have raised the share of exits from
unemployment to permanent employment in total exits from unemployment by about 14%, taking the point
estimates at face value.
Results from transitions from unemployment to permanent employment also indicate a greater impact
of the reform in small establishments than in larger ones (Table A4). The results from the estimation of the
competing-risk hazard models suggest that the reform increased the average transition to a permanent
contract – conditional on being unemployment for six months or less – by 26% in establishments with
50 or fewer employees, compared with 15% in establishments with over 50 employees (Figure A6). To the
extent that the large majority of firms are mono-establishments, this points to a greater impact of the
reform on hiring for smaller firms.
80. As for analyses based on the ECL, establishment here refer to the Cuenta de Cotización, which is unique
for each firm at the province level.
81. However, in absolute terms, the estimated effect on transitions to permanent contracts is no smaller than
that on transitions to temporary contracts. For example, taking point estimates at face value, in the first 6
months of unemployment, the average probability of a monthly transition to a temporary job increased
from 18.6% to 19.9% due to the reform whereas the average transition to permanent employment increased
only from 1.7% to 2.1%.
![Page 71: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/71.jpg)
70 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Table A9. The effect of the reform on Transitions out of Unemployment into temporary employment and into permanent employment
Monthly data, coefficients of the competing-risk hazard models
Post-reform dummy
Placebo tests Number of observations
Years
One quarter before reform
Two quarters before reform
Three quarters before reform
(1) (2) (3) (4)
All sample U to TC 0.0999 -0.120* -0.177*** -0.149***
3,886,004 2003-2012 (0.0641) (0.0654) (0.0470) (0.0442)
U to PC 0.256*** -0.0836 -0.175** -0.113*
(0.0746) (0.0866) (0.0705) (0.0630)
Establishments with 50 employees or less
U to TC 0.0967 0.0994* 0.0377 0.0254
2,174,106 2006-2012
(0.0669) (0.0541) (0.0576) (0.0470)
U to PC 0.264*** 0.0381 -0.0659 -0.0264
(0.0796) (0.0789) (0.0718) (0.0674)
Establishments with more than 50 employees
U to TC 0.0947* 0.0574 0.0117 -0.00221
(0.0514) (0.0574) (0.0690) (0.0473)
U to PC 0.165*** -0.0401 -0.0290 0.205***
(0.0629) (0.0965) (0.114) (0.0662)
Note: U: unemployment; PC: permanent employment; TC: temporary employment. Coefficients within border lines refer to simultaneous estimates. The All equations control for the unemployment duration (in logs) up to the 3rd order, the national unemployment rate, change in regional employment (and its interaction with the unemployment duration (in logs), a 3rd order polynomial in time (months) in the models of the first two lines and a 5th order polynomial in the others, age and education categories, an indicator if the individual receives unemployment benefits (and the interaction of those variables with the duration of the unemployment spell), gender, migrant status, region, characteristics of the previous job (sector, type of contract, and if the individual was dismissed), the number of times the individual has been unemployed, the percentage of time he has been employed as well as monthly dummies. The sample includes all individuals who have been unemployed at least one day. Unemployment durations are censored at 30 months in unemployment. Robust standard errors, clustered on time in parentheses. ***, **, *: significant at the 1%, 5% and 10% level, respectively.
Figure A5. Exit from unemployment by type of contract and unemployment duration
Fitted exit probabilities, by calendar month
Note: The figure presents the average of estimated monthly probabilities of individual transitions from unemployment to employment by type of contract in the new job and unemployment duration (less than six months, from 7 to 12 months and more than 12 months), netting out the effect of the time trend and monthly dummies. The sample includes all individuals who have been unemployed at least one day. Unemployment durations are censored at 30 months in unemployment. Panel A refers to exits from unemployment to temporary contracts and Panel B to exits from unemployment to permanent contracts.
0
0.05
0.1
0.15
0.2
0.25
200
6m1
200
6m5
200
6m9
200
7m1
200
7m5
200
7m9
200
8m1
200
8m5
200
8m9
200
9m1
200
9m5
200
9m9
201
0m1
201
0m5
201
0m9
201
1m1
201
1m5
201
1m9
201
2m1
201
2m5
201
2m9
Less than 6 months From 7 to 12 months More than 12 months
0
0.005
0.01
0.015
0.02
0.025
200
6m1
200
6m5
200
6m9
200
7m1
200
7m5
200
7m9
200
8m1
200
8m5
200
8m9
200
9m1
200
9m5
200
9m9
201
0m1
201
0m5
201
0m9
201
1m1
201
1m5
201
1m9
201
2m1
201
2m5
201
2m9
A. Exit to temporary contract B. Exit to permanent contract
![Page 72: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/72.jpg)
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT – 71
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Figure A6. Exit from unemployment conditional on duration smaller or equal than 6 months, by type of contract and establishment size
Fitted exit probabilities, by calendar month
Note: The figure presents the average of estimated monthly probabilities of individual transitions from unemployment to employment by type of contract in the new job and unemployment duration (less than six months, from 7 to 12 months and more than 12 months), netting out the effect of the time trend and monthly dummies. The sample includes all individuals who have been unemployed at least one day. Unemployment durations are censored at 30 months in unemployment. Panel A refers to exits from unemployment to temporary contracts and Panel B to exits from unemployment to permanent contracts.
Placebo tests are run for all the estimations by “anticipating” the date of the reform by up to three
quarters. The coefficients are either non-significant or negative for all specifications, except for the placebo
test three quarters before the reform for transitions to permanent employment in establishments with more
than 50 employees. Similarly, some caution is required in interpreting baseline estimates since placebo test
statistics are significant (albeit negative) when the placebo date is two quarters before the true reform.
However, these statistics turn out to be insignificant when transition hazards are allowed differing across
establishment types.
Several robustness checks have been performed as well, which include i) substituting the polynomial
trend for calendar year dummies; ii) using polynomial time trends of different order; ii) performing the
estimations separately by type of previous contract hold, by age groups and by gender; iii) substituting
regional unemployment rates for aggregate ones; and iv) replacing registered employment by the index of
economic activity of FEDEA. Results presented in Table A9 are robust to these changes.
The effect of the 2012 reform has not been homogeneous on all individuals (Table A10). In particular,
transitions from unemployment towards permanent employment have increased significantly more for
those that had a temporary contract prior to becoming unemployed, while the effect remains insignificant
for those who previously held a permanent contract. Symmetrically, the results suggest a marginally
positive impact of the reform on the speed of transition towards a temporary contract only for those that
had a permanent contract prior to becoming unemployed.
The reform seems also to have had a greater impact on younger workers, a population group severely
affected by unemployment, than on older ones. Transition rates from unemployment towards permanent
employment have increased more for individuals aged 18-30, compared to for those aged 31-44 and 45-64
(Table 6).
Finally, estimating the same models for transitions out of permanent or temporary contract yields an
insignificant impact of the reform on these transition hazards.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
200
6m1
200
6m5
200
6m9
200
7m1
200
7m5
200
7m9
200
8m1
200
8m5
200
8m9
200
9m1
200
9m5
200
9m9
201
0m1
201
0m5
201
0m9
201
1m1
201
1m5
201
1m9
201
2m1
201
2m5
201
2m9
Firm size <= 50 Firm size > 50
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
200
6m1
200
6m5
200
6m9
200
7m1
200
7m5
200
7m9
200
8m1
200
8m5
200
8m9
200
9m1
200
9m5
200
9m9
201
0m1
201
0m5
201
0m9
201
1m1
201
1m5
201
1m9
201
2m1
201
2m5
201
2m9
A. Exit to temporary contract B. Exit to permanent contract
![Page 73: The 2012 Labour Market Reform in Spain: A Preliminary Assessment](https://reader034.vdocument.in/reader034/viewer/2022052419/589d9f271a28abd6498bd023/html5/thumbnails/73.jpg)
72 – THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT
THE 2012 LABOUR MARKET REFORM IN SPAIN: A PRELIMINARY ASSESSMENT DECEMBER 2013 ©OECD 2013
Table A10. The effect of the reform on Transitions out of Unemployment into temporary employment and into permanent employment, by personal characteristics
Monthly data, coefficients of the competing-risk hazard models
Post-reform dummy Number of observations
U to TC U to PC
(1) (2) (3)
Previous temporary 0.0902 0.639*** 3,047,292
(0.0637) (0.0775)
Previous permanent 0.149* 0.153 838,712
(0.0822) (0.0993)
Males 0.104 0.269*** 1,996,002
(0.0682) (0.0663)
Females 0.0913 0.246*** 1,890,002
(0.0606) (0.0873)
Age 18-30 0.0703 0.319*** 1,770,386
(0.0542) (0.0544)
Age 31-44 0.0950 0.247*** 1,370,296
(0.0702) (0.0836)
Age 45-64 0.191* 0.225* 745,322
(0.109) (0.124)
Note: U: unemployment; PC: permanent employment; TC: temporary employment. Coefficients on the same line refer to simultaneous estimates. All equations control for the unemployment duration (in logs) up to the 3rd order, the national unemployment rate, change in regional employment (and its interaction with the unemployment duration (in logs), a 3rd-order polynomial in time (months), age and education categories, an indicator if the individual receives unemployment benefits (and the interaction of those variables with the duration of the unemployment spell), gender, migrant status, region, characteristics of the previous job (sector, type of contract, and if the individual was dismissed), the number of times the individual has been unemployed, the percentage of time he has been employed as well as monthly dummies. The sample includes all individuals who have been unemployed at least one day. Unemployment durations are censored at 30 months in unemployment. Robust standard errors, clustered on time in parentheses. ***, **, *: significant at the 1%, 5% and 10% level, respectively.